abstractcount.txt

present technique memory time volume rendering curvilinear adaptive mesh refinement data defined within extrudable computational space one main challenge ray casting curvilinear volume linear viewing ray physical space typically correspond curved ray computational space proposed method utilizes specialized representation curvilinear space provides compact representation parameter transformation computational space physical space without requiring extensive preprocessing simplifying representation computational space position using extrusion profile surface requisite transformation greatly simplified implementation achieves interactive rate minimal load time memory overhead using commodity graphic hardware real data field volume rendering pre arbitrary transfer function certainly led significant convincing result quality performance wise allowing high quality visualization standard pc consumer graphic showing ideal scalar signal along cast ray better approximated succession polynomial curve opposed linear segment propose new method pre volume rendering method based second order polynomial interpolation scalar value allowing converge rapidly towards integration volume reconstructed trilinear filter approach manages capture smoothness volume detail without need ray resampling consequently succeeds reducing visual artefact comparison previous technique furthermore adapt existing technique compute pre table using gpu thus making approach suitable transfer function manipulation volume rendering multiple intersecting volumetric object difficult visualization task especially different rendering style need applied component order achieve desired illustration effect real performance even complex scenario obtained exploiting speed flexibility modern gpus time programming necessary shaders turned task gpu expert foresee demand intermediate level programming abstraction visualization specialist realize advanced application without need deal shader programming intricacy paper describe generic technique multi rendering generates shader code dynamically abstract render graph combining pre node complex volume operation realized system efficiently creates gpu fragment shader vertex shader program achieve desired visual result demonstrate flexibility technique applying several dynamically generated volume rendering style multi medical datasets present non realistic rendering technique interactive exploration isosurfaces generated remote volumetric data instead relying conventional smooth shading technique render isosurfaces point technique used represent render isosurfaces remote client environment non realistic nature proposed rendering method enables server transmit essential surface feature substantially reduces network traffic algorithm also utilizes frame coherence efficiently encodes isosurface configuration inside voxel cell minimize network overhead finally algorithm adjust point distribution using different illumination setting adapt different network speed define concept crossing number simultaneous graph extending crossing number problem traditional graph discus difference traditional crossing number problem give np proof lower upper bound new problem furthermore show existing heuristic exact algorithm traditional problem adapted new task simultaneous crossing minimization report brief experimental study implementation visualizing graph node diagram commonly accepted employed general rule number link crossing minimized whenever possible however little attention paid handle remaining crossing visualization study presented paper examines effect crossing angle performance path tracing task found effect varied size crossing angle particular task response time decreased crossing angle increased however rate decrease tended level angle close degree one implication study graph visualization minimizing crossing number sufficient reduce negative impact minimum angle remaining crossing maximized well propose method visually summarize collection network clustering vertex given method allows efficient comparison individual network well visualizing average composition structure set network concrete application analyze set several hundred personal network migrant individual level network image provide visual hint assessing mode acculturation respondent population level show cultural integration varies specific characteristic migrant country origin year residence skin color effectively visualizing complex node graph depict relationship among data node challenging task due clutter occlusion resulting excessive amount edge paper propose novel energy hierarchical edge clustering method node graph taking consideration graph topology method first sample graph edge segment using delaunay triangulation generate control point hierarchically clustered energy optimization edge grouped according position direction improve comprehensibility abstraction thus reduce visual clutter experimental result demonstrate effectiveness proposed method clustering edge providing good high level abstraction complex graph modern experimental computational fluid mechanic increasingly concerned structure nature fluid motion recent research highlighted analysis one transport structure called lagrangian coherent structure however quantity nature flow transport still unclear paper focus transport characteristic physical quantity propose approach visualize finite transport structure quantity advection similar integral convolution scalar field along path flow field applied well set physical quantity field yield structure giving insight dynamical process underlying flow demonstrate approach number test data set particle tracing established powerful visualization technique show dynamic flow particle tracing however quickly overextends viewer due massive amount visual information typically produced technique paper present strategy reduce amount time revealing important structure flow importance measure introduce simple yet effective clustering approach vector field use scalar flow quantity different scale combination user region interest measure used control shape appearance density particle way user focus dynamic important region time preserving context information also introduce new focus particle tracing called anchor line anchor line used analyze local flow feature visualizing much particle separate time long take separated fixed distance particular interest finite time lyapunov exponent scalar quantity measure rate separation infinitesimally close particle flow used guide placement anchor line effectiveness approach visualization flow field validated using synthetic field well real simulation data inspired abstracting focusing explanatory quality diagram drawing art paper propose novel seeding strategy generate representative illustrative streamlines vector field enforce visual clarity evidence particular focus algorithm depict underlying flow pattern effectively succinctly minimum set streamlines achieve goal distance field generated encode distance grid point field nearby streamlines local metric derived measure dissimilarity vector original field approximate field computed distance field global metric used measure dissimilarity streamlines based local error decide whether drop new seed local point process iterated generate streamlines streamlines found dissimilar existing one present example image generated algorithm report result qualitative analysis user study parallel performance challenging topic streamline visualization large unstructured flow datasets parallel distributed computer depends strongly domain partition unsuitable partition often lead severe load imbalance high frequent communication among domain partition address problem present approach flow data partitioning taking account flow direction feature multilevel spectral graph bisection method employed reduce communication synchronization overhead among distributed domain edge weight corresponding adjacent matrix defined based anisotropic local diffusion operator assigns strong coupling along flow direction weak coupling orthogonal flow meanwhile distribution seed point flow feature vortex structure also considered partitioning obtain good load balance experimental result given show feasibility effectiveness method present new method visual analysis multivariate system trace method combine three perspective schematic diagram time series plot state transition graph show perspective related discus integration single solution visual analysis multivariate system trace combination three perspective provides user rich analysis interface enables gaining significant insight system behavior illustrate advantage approach providing real use case facial emotion expressive facial motion become intrinsic part many graphic system human computer interaction application dynamic high dimensionality facial motion data make exploration processing challenging paper propose novel visualization system expressive facial motion data exploration based principal component analysis dimensionality reduction anatomical facial sub region high dimensional facial motion data mapped space rendered colored trajectory color represents different emotion design intuitive interface allow user effectively explore analyze high dimensional facial motion space application visualization system novel facial motion synthesis emotion recognition demonstrated event system gaining increasing popularity building loosely coupled distributed system since business process becoming interconnected event event system fit well supporting monitoring business process paper present event business intelligence tool event tunnel framework provides interactive visualization event stream support business analyst exploring business incident visualization based metaphor considering event stream cylindrical tunnel presented user multiple perspective information single event laid event tunnel encoded event glyph allow selective mapping event attribute color size position different policy placement event tunnel well clustering mechanism generate various view historical event data event tunnel able display relationship event facilitates user discover root cause causal dependency event pattern framework couple event visualization query tool allow user search relevant event within data repository using query filler highlighting operation analyst navigate event tunnel required information event pattern become visible demonstrate approach use case fraud management logistics domain motivated growing interest use ridge scientific visualization analyze two height ridge definition eberly lindeberg propose raw feature definition leading superset ridge point obtained two definition set raw feature point correct dimensionality narrowed either eberly lindeberg ridge using boolean filter formulate straight computation height ridge requires explicit eigenvalue calculation avoided using equivalent definition raw feature set give derivation describe efficient algorithm two special case height ridge dimension one co one alternative aforementioned filter propose new criterion filtering raw feature based distance contour generally make better decision demonstrate synthetic field topographical dataset fluid flow simulation dataset set test data show unavoidable use filter eliminate false positive purpose use angle feature tangent slope line quality measure based formalize previously published filter adaptive filtering method volume data presented paper filtering method input data set create hierarchy multiple data set data classification task performed level data pyramid decide local structure type data voxels classified linear planar blob structure based gradient eigenvalue hessian matrix classification result used adjust shape orientation filter noise suppressed key feature preserved present new framework combine hierarchical multi representation video compression manage render large scale time data preprocessing step proposed method first construct multi hierarchy using octree structure individual time step applies motion prediction compress octree node rendering stage data decompressed rendered using hardware texture mapping proposed approach eliminates hierarchical decompression dependency commonly found conventional hierarchical wavelet representation method lead efficient reconstruction data along time axis system provides user spatial region adjust spatial level selection temporal roi sub frequent update playback suitable control roi system reach interactive playback frame rate allows user observe dynamic nature large time data set large amount time datasets create great challenge user understand explore paper proposes efficient visualization method observing overall data content change throughout entire time dataset develop interactive storyboard approach composing sample volume rendering descriptive geometric primitive generated data analysis process storyboard system integrates automatic visualization generation method interactive adjustment procedure provide new tool visualizing exploring time datasets also provide flexible framework quantify data difference automatically select representative datasets exploring scientific data distribution feature since approach reduces visualized data amount understandable size format user used effectively visualize represent explore large time dataset initial user study result show approach shortens exploration time reduces number datasets user visualized individually visualization method especially useful situation require close observance capable interactive rendering documentation demonstration radial space visualization useful representing distribution attribute hierarchical data however also suffers drawback term view transition context preservation thin slice flexibility large sized data support address problem propose fanlens enhancement upon existing approach new feature like incremental layout fisheye distortion based selecting visual toolkit also feature dynamic hierarchy specification dynamic visual property mapping smooth animation etc illustrate effectiveness technique two example case study result informal user experiment animation frequently utilized visually depict change time varying data set task natural fit yet explicit animation rarely employed static data discus use animation overcome three common limitation faced information visualization application context small device constraint output display limited interaction possibility high data density provide concrete example applying animation combat limitation four common visualization type geospatial data treemaps parallel coordinate display large graph unlike previous work examines animation maintaining user orientation view change displaying data variable discus animation utility multiplexing available screen space context constrained display demonstrate ability effectively gain screen resolution quickly uncover trend help find unexpected data pattern reduce visual clutter paper describes whatsonweb search clustering engine allows user browse analyze result query mean enhanced graph visualization technique whatsonweb integrates wide array visual interface animation interaction functionality clustering technology effectiveness different visual interface different clustering algorithm implemented system measured mean extensive experimental analysis described system represents significant evolution previous clustering engine web widespread use mobile device brings opportunity capture large continuous information human behavior mobile data tremendous value leading business opportunity market strategy security concern etc visual analytics system support interactive exploration discovery needed extracting insight data however visual analysis complex social mobile data present several challenge created mobivis visual analytics tool incorporates idea presenting social spatial information one heterogeneous network system support temporal semantic filtering interactive time chart ontology graph respectively data subset interest isolated close investigation ring compact radial representation individual group behavior introduced allow easy comparison behavior pattern demonstrate capability mobivis result obtained analyzing mit reality mining dataset paper present interactive visualization toolkit navigating analyzing national science foundation funding information design build upon treemap layout stacked graph contribute customized technique visually navigating interacting hierarchical data nsf program proposal supporting visual search analysis allowing user make informed decision visualization toolkit propose two visualization technique simplify navigation hierarchical data dimensional treemaps make hierarchical structure easily recognized labeled treemap help user get clear overview content structure make internal area rectangle correspond weight data set furthermore incremental layout method adopted handle information large scale improved treemap visualization help visually analyze static funding data stacked graph utilized analyze time data visual analysis technique research trend nsf popular nsf program quickly identified primary contribution demonstration novel way effectively present analyze nsf funding data success open source software project apache mozilla come opportunity study development process paper present stargate novel system visualizing software project whereas previous software project visualization concentrated mainly source code change literally place developer center design developer grouped visually cluster corresponding area file repository work connection drawn people communicate via email change repository also displayed stargate easy look beyond source code see trend developer activity system used anyone interested project especially benefit project manager project novice software engineering researcher augment schematic map transportation system superimposing street map fitted using image warping technique schematic map successful conveying information line connection public transportation network usually contain little detail describing environment station embedding surrounding area annotation distorted city map therefore alleviates deficiency improves usability schematic transportation map merging two different navigational space technique fitting street map schematic map based moving least square combination overlap control technique thus obtain easily readable transportation network map show typical city map feature river street park without compromising schematization furthermore interactive exploration couple zooming warping control level detail call warping zoom though medium visualization limited potential dimension dataset many area scientific study understanding correlation dimension uncertainty pivotal mining useful information dataset obtaining insight necessitate visualizing many relationship among temporal spatial dimensionality data uncertainty utilize multiple view interactive dataset exploration selection important feature apply technique unique challenge cosmological particle datasets show interactivity incorporation multiple visualization technique help overcome problem limited visualization dimension allow many type uncertainty seen correlation variable present zoomable adjacency matrix explorer visualization tool exploring graph scale million node edge zame based adjacency matrix graph representation aggregated multiple scale allows analyst explore graph many level zooming panning interactive performance overview detailed view several component work together zame tool make possible efficient matrix ordering algorithm group related element individual data case aggregated higher meta aggregate arranged pyramid hierarchy allows paging gpu shader program support smooth multiscale browsing using zame able explore entire french wikipedia article link interactive performance standard consumer computer hardware multivariate network data set describe relationship set entity also attribute paper present new technique determine layout multivariate network using geodesic self map training process geosom graph distance non combined attribute similarity based network graph distance distribution resulted layout le edge crossing generated previous method conducted user study evaluate effectiveness hybrid approach result compared commonly used glyph technique user study show hybrid approach help user draw conclusion relationship vertex attribute multivariate network quickly accurately addition user found easier compare different relationship set entity finally capability hybrid approach demonstrated using world military expenditure weapon transfer network abstract graph network commonly recurring data type many field order visualize graph effectively graph must laid screen coherently many algorithm exist many algorithm tend slow input graph large paper present new approach large graph layout problem quickly generates effective layout new method proceeds generating clustering hierarchy graph applying treemap hierarchy finally placing graph vertex associated region treemap ideal interactive system operation semantic zooming performed since work done initial hierarchy calculation take little work recalculate layout method also valuable resulting layout used input iterative algorithm force directed method greatly reduces number iteration required converge near optimal layout analyze time data set tracking feature time often necessary better understand dynamic nature underlying physical process tracking time feature however non boundary feature cannot easily defined paper propose new framework visualize time feature motion without explicit feature segmentation tracking framework time feature described time series time activity curve compute distance similarity voxel time series feature use dynamic time warping distance metric purpose dtw compare shape similarity two time series optimal warping time phase shift feature time accounted applying dtw compare voxel time series feature time distance field computed amount time warping required voxel match feature provides estimate time feature likely occur based tac distance field several visualization method derived highlight position motion feature present several case study demonstrate compare effectiveness framework work introduces novel streamline seeding technique based dual streamlines orthogonal vector field instead tangential greedy algorithm presented produce net orthogonal streamlines iteratively refined resulting good domain coverage high degree continuity uniformity algorithm easy implement efficient naturally extends curved surface ability extract follow time flow feature volume data generated large numerical simulation enables scientist effectively see validate modeled phenomenon process extracted feature often take much le storage space computing resource visualize feature extraction tracking method first identify feature interest time step independently correspond feature consecutive time step data since method handle time step separately use coherency feature along time dimension extraction process paper present prediction method us prediction step make best guess feature region subsequent time step followed growing shrinking border predicted region coherently extract actual feature interest method make use temporal coherency data accelerate extraction process implicitly solving tedious correspondence problem previous method focus method low cost little storage overhead thus facilitates interactive runtime extraction visualization unlike previous method largely suited batch processing due high computational cost visualization actual condition organization challenging issue propose new system called business microscope sens activity people organization provides visual feedback user use name shaped sensor node measure face interaction employee massive amount data collected sensor terminal signal server displayed organization topographic map display frequency organizational activity depict organization topographic map system creates novel relation tree using interaction data pair member kind map group organization hierarchically form island member island relationship many others form mountain plotted contour comprehend actual condition organization topographic map tested technique several experiment evaluate system present new method combined visualization software architecture diagram uml class diagram component diagram software metric defined group diagram element method extends existing rendering technique area interest system architecture diagram visualize several metric possibly missing value defined overlapping area interest use solution combine texturing blending smooth scattered point interpolation new method simplifies task visually correlating distribution outlier value multivariate metric dataset system structure demonstrate application method component class diagram extracted real system visualizing large online social network challenging yet essential task paper present himap system visualizes clustered graph via hierarchical grouping summarization himap employ novel adaptive data loading technique accurately control visual density graph view along optimized layout algorithm two kind edge bundling method effectively avoid visual clutter commonly found previous social network visualization tool himap also provides integrated suite interaction allow user easily navigate social map smooth coherent view transition keep momentum finally confirm effectiveness himap algorithm graph based evaluation although significant progress made toward effective insight discovery visual sense making approach lack effective efficient approach manage large amount insight discovered paper propose systematic approach leverage problem around concept fact fact refer pattern relationship anomaly extracted data analysis direct product visual exploration permit construction insight together user mental model evaluation different mental model type fact discovered data predictable application thus possible develop general fact management framework allow visualization user effectively efficiently annotate browse retrieve associate exchange fact since fact essential component insight feasible extend fmf effective insight management variety visual analytics approach toward goal first construct fact taxonomy categorizes various fact multidimensional data capture essential attribute extensive literature survey user study propose conceptual framework fact management based upon fact taxonomy concrete scenario visual sense making real data set illustrates fmf work current visualization system typically based concept interactive post decoupling data visualization process data generation offer flexible application visualization tool also lead however information loss visualization therefore combination visualization data generating process visualization produced data offer significant support understanding abstract data set well underlying process due application characteristic data generating process task requires tailored visualization concept work focus application field simulating biochemical reaction network discrete system stochastic process generate multi multivariate time analyzed compared three different process level model experiment level multi simulation data associated broad range analysis goal meet challenging characteristic present visualization concept specifically tailored three process level fundament three visualization concept compact view relates multi simulation data characteristic model structure experiment view provides visualization experiment level visualization model level coordinate multiple instance view comparison experiment level multi simulation data view give overview data analyzed detail time view suited analysis goal although derive visualization concept one concrete simulation process general concept tailoring visualization concept process level generally applicable visualization simulation process particle simulation popular tool researcher various science combination availability ever larger cot cluster consequently increasing number simulated particle resulting datasets pose challenge real visualization additionally semantic density particle exceeds possibility basic glyph like splat sphere result dataset size larger least order magnitude interactive visualization common workstation requires careful optimization data management especially transfer cpu gpu propose flexible benchmarking tool along series test allow evaluation performance different cpu combination relation particular implementation evaluate different uploading strategy rendering method point compound glyph suitable representing aforementioned datasets cpu gpu approach compared respect rendering storage efficiency point optimal solution dealing time datasets result research general interest since transferred application cpu bandwidth high number graphical primitive per dataset pose problem employed tool set streamlining measurement process made publicly available paper propose novel volume rendering algorithm large time field exploring temporal spatial coherence important direction speeding rendering time data previously technique hierarchically partition time space domain data structure result previous time step multiresolution rendering however studied domain partitioned first obtain better rate address open question show theoretically experimentally partitioning time domain first better call resulting structure binary time tree primary structure octree secondary structure space time tree typically spt rendering higher level detail higher rate run faster addition devise novel cut algorithm facilitate efficient volume rendering using spt tree develop novel preprocessing algorithm build spt tree propose modified error metric theoretical guarantee monotonicity property desirable tree search experiment datasets large gb using pc gb ram demonstrated efficacy new approach space layout technique tree representation frequently used available screen space small data set large paper propose new approach space tree representation us mechanism point rendering paradigm additionally helpful interaction technique tie layout presented relate new technique established space technique along line newly developed classification also evaluate numerically using measure ink overplotted graph data mining algorithm rely graph canonical form compare different graph structure canonical form definition depend node edge label paper introduce unique canonical visual matrix representation depends graph topological information two structurally identical graph exactly visual adjacency matrix representation canonical matrix node ordered based breadth search spanning tree special rule filter designed guarantee uniqueness arrangement unique matrix representation provides persistence stability used harnessed visualization especially data exploration study symbol frequently used represent data object visualization appropriate contrast symbol precondition determines efficiency visual analysis process study contrast different type symbol context scatterplots based user testing quantitative model symbol contrast total different symbol generated using four size two class asterisk shaped four category rotational symmetry used three different task user test result internal separation space established symbol type study space every symbol represented point visual contrast defined task performance symbol represented distance point position point space obtained multidimensional scaling reveal effect different visual feature scale also larger distance imply better symbol separation visual task therefore indicate appropriate choice symbol resulting configuration discussed number pattern relation property symbol resulting contrast identified short found size effect space linear dominant shape effect paper present novel adaptive visualization technique constituting polygon dynamically change geometry visual attribute depending user interaction change take place objective conveying required level detail user view proposed technique successfully applied build treemap visualization hierarchical data widely used information structure new visualization exploit adaptive nature address issue like cluttered display imperceptible hierarchy lack smooth zoom technique common tree visualization also present algorithm utilizes flexibility proposed technique deal occlusion problem inherent information visualization one hand work establishes adaptive visualization mean displaying tree data promotes technique potential candidate employed visualize information structure also many graph visualization system use graph hierarchy organize large input graph logical component approach detect feature globally data place feature inside level hierarchy however feature detection global process consider node graph near feature interest tuggraph system exploring path proximity around node subgraphs graph approach modifies pre hierarchy order see node subgraph interest extends larger graph guaranteed create path hierarchy abstraction shown meaningful respect structure graph system work well graph hundred thousand node million edge tuggraph able present view proximal information context entire graph second require layout full graph input many graph used real application consist node belonging one category call graph graph social network typical example multiple graph node person link friendship category community person belong often helpful visualize connectivity category graph simultaneously paper present new visualization technique multiple graph technique firstly construct hierarchical cluster node based connectivity category place node new hybrid space force layout algorithm clearly display connectivity category information show layout result using hybrid method compare method present case study using active biological network dataset spring model based force directed algorithm widely used drawing undirected graph sophisticated implementation efficient visualizing large graph however practical experience show many case layout quality suffers result non vertex density give rise warping effect vertex outskirt drawing often closer near center branch tree graph tend cling together paper propose algorithm overcome effect algorithm combine efficiency good global structure spring model flexibility kamada stress model specifying ideal edge length effective overcoming warping effect end result graph visualization people read graph understand data make effective essential construct visualization based people read graph despite popularity importance graph usage variety application domain little known people read graph lack knowledge severely limited effectiveness graph visualization attempt understand people read graph previously observed people geodesic tendency based subjective eye tracking data paper present two controlled experiment one approve existence geodesic tendency examine effect tendency people reading graph result show performing path search task eye encounter node one link link go toward target node likely searched first result also indicate graph drawn branch link path leading away target node graph reading performance significantly improved diffusion tensor field reveal underlying anatomical structure biological tissue neural fiber brain current method visualizing diffusion tensor field categorized two class integral curve glyph integral curve continuous represent underlying fiber structure prone integration error loss local information glyph useful representing local tensor information convey connectivity anatomical structure well introduce simple yet effective visualization technique extends streamball method flow visualization tensor ellipsoid tensor ellipsoid represents local tensor either blend neighboring tensor break away depending orientation anisotropy resulting visualization show connectivity information underlying anatomy characterizing local tenor detail interactively changing iso parameter user examine diffusion tensor field entire spectrum continuous integral curve discrete glyph expert evaluation indicates method conveys useful visual information local anisotropy white matter fiber information previously unavailable tractography model method provides visual tool assessing variability dti fiber tract integrity relation function high angular resolution diffusion imaging emerging magnetic resonance imaging technique overcomes decisive limitation predecessor diffusion tensor imaging hardi resolve locally one direction diffusion pattern water molecule thereby open opportunity display track crossing fiber showing local structure reconstructed angular probability profile fast detailed interactive way improve quality research area help move clinical application paper present novel approach hardi glyph visualization generally visualization function resides sphere expressed laplace series gpu glyph rendering improves performance traditional way hardi glyph visualization well visual quality reconstructed data thus offering interactive hardi data exploration local structure white brain matter paper exploit capability modern gpus overcome large processor memory data visualization present correlation study time multivariate volumetric data set scientific discipline test hypothesis discover insight scientist interested looking connection among different variable among different spatial location within data field response propose suite technique analyze correlation time multivariate data various temporal curve utilized organize data capture temporal behavior reveal pattern find connection perform data clustering segmentation using k clustering graph partitioning algorithm study correlation structure single pair variable using pointwise correlation coefficient canonical correlation analysis demonstrate approach using result time multivariate climate data set paper propose stochastic simulation model analyze cellular signal transduction high number object simulation requires advanced visualization technique first handle large data set second support human perception crowded environment third provide interactive exploration tool adjust state cell external signal specific set signaling molecule transport information nucleus deep inside cell key molecule regulate gene expression contrast continuous ode model model signaling molecule individually realistic crowded disordered environment beyond spatiotemporal concentration profile data describes process mesoscopic molecular level allowing detailed view intracellular event proposed schematic visualization individual molecule track reaction selected brought focus highlight signal transduction pathway segmentation depth cue depth field applied reduce visual complexity also provide virtual microscope display image comparison wet lab experiment method applied distinguish different transport mode mapk protein kinase signaling molecule cell addition simulate diffusion drug molecule extracellular space solid tumor visualize challenge cancer related therapeutic drug delivery direct volume rendering important tool visualizing complex data set however process generating image data information lost form attenuation occlusion lack feedback mechanism quantify loss information rendering process make design good transfer function difficult time consuming task paper present notion visibility transfer function transfer function provide good visibility feature interest given viewpoint achieve introduce visibility histogram histogram provide graphical cue intuitively inform user contribution particular scalar value final image carefully manipulating parameter opacity transfer function user maximize visibility interval interest volume data set based observation also propose semi method generating transfer function progressively improves transfer function defined user according certain importance metric user deal tedious task making small change transfer function parameter rely system perform search automatically methodology easily deployed visualization system used together traditional opacity transfer function based scalar value well multidimensional transfer function sophisticated rendering algorithm paper present novel method interactive identification contextual interest point within volumetric data picking direct volume rendered image clinical diagnostics point interest often located center anatomical structure order derive volumetric position allows convenient examination intended structure system automatically extract contextual meta information dicom imaging communication medicine image setup medical workstation along viewing ray volumetric picking ray profile analyzed structure similar predefined template knowledge base demonstrate result obtained position utilized highlight structure slice view interactively calculate centerline tubular object place label contextually volumetric position viewpoint selection becoming useful part volume visualization pipeline improves efficiency data understanding providing representative viewpoint present two structure view descriptor shape view descriptor detail view descriptor select optimal viewpoint maximum amount structural information two proposed structure view descriptor based gradient direction gradient well measurement boundary structure proved feature interest many application shape view descriptor designed evaluate overall orientation feature interest estimating local detail employ bilateral filter construct shape volume bilateral filter effective smoothing local detail preserving strong boundary structure time therefore large global structure shape volume small local detail still remain original volume detail view descriptor measure amount visible detail boundary structure term variance local structure shape volume original volume two view descriptor integrated viewpoint selection framework framework emphasize global structure local detail flexibility tailored user specific situation performed experiment various type volume datasets experiment verify effectiveness proposed view descriptor proposed viewpoint selection framework actually locates optimal viewpoint show maximum amount structural information define transfer function based first second statistical moment consider evolution mean variance respect growing neighborhood around voxel evolution defines curve identify important trend project back resulting projection brushed easy robust classification material material border transfer function applied ct mr data rapid development graphic hardware volume rendering technique many volumetric datasets rendered real time standard pc equipped commodity graphic board however effectiveness result especially direct volume rendered image difficult validate user may aware ambiguous even misleading information result limit application volume visualization paper introduce four quantitative effectiveness measure distinguishability contour clarity edge consistency depth coherence measure target different effectiveness issue direct volume rendered image based measure develop visualization system automatic effectiveness assessment providing user instant feedback effectiveness result case study user evaluation demonstrated high potential system present novel classification technique volume visualization take shape volumetric feature account presented technique enables user distinguish feature based shape assign individual optical property based rough pre done windowing exploit curve volumetric structure order derive shape descriptor similar used current shape recognition algorithm shape descriptor distinguishes three main shape class longitudinal surface blobby shape contrast previous approach classification performed per level assigns uniform shape descriptor feature therefore allows intuitive user interface assignment optical property using proposed technique becomes instance possible distinguish blobby heart structure filled contrast agent potentially occluding vessel rib bone introducing basic concept show presented technique performs real world data discus current limitation difficult task design transfer function noisy data traditional transfer space data value different material overlap paper introduce novel statistical transfer space presence noise separate different material volume data set method adaptively estimate statistical property mean value standard deviation data value neighborhood sample point property used define transfer space enables distinction different material additionally present novel approach interacting new transfer space enables design transfer function based statistical property furthermore demonstrate statistical information applied enhance visual appearance rendering process compare new method lh transfer function demonstrate usefulness paper present interactive transfer function design tool based ellipsoidal gaussian transfer function approach explores volumetric feature statistical space modeling space using gaussian mixture model small number gaussians maximize likelihood feature separation instant visual feedback possible mapping gaussians etf analytically integrating etf context pre volume rendering process suite intuitive control widget designed offer automatic transfer function generation flexible manipulation allowing inexperienced user easily explore undiscovered feature several simple interaction gpu implementation demonstrates interactive performance plausible scalability compare favorably existing solution effectiveness approach verified several datasets information visualization show tremendous potential helping expert casual user alike make sense temporal data current time series visualization tool provide poor support comparing several focus temporal dataset retaining context distance awareness introduce method supporting kind multi interaction call stack zooming approach based user interactively building hierarchy strip stacked top subsequent stack represents higher zoom level sibling strip represent branch visual exploration correlation graphic show relation stack strip different level providing context distance awareness among focus point zoom hierarchy also used graphical history communicating insight stakeholder also discus visual space support stack zooming extended annotation local statistic computation fit hierarchical stacking metaphor present prototype interface visualizing interacting univariate time digital signal kind signal might measured using oscilloscope analyzed find anomaly edge timing overshoot measured characteristic prototype visualizes data using standard eye diagram wrap signal time well parallel coordinate view novel layer technique based drawer metaphor spreading eye diagram apart de trace within opened trace within sorted grouped according various attribute drilling data achieved opening sub also present novel technique browsing eye diagram called temporal brushing initial user feedback student electrical engineering reported video demonstrating prototype downloaded http wikipedia become one used knowledge base worldwide problem trustworthiness information disseminates becomes central wikipediaviz introduce five visual indicator integrated wikipedia layout keep casual wikipedia reader aware important metainformation article read design wikipediaviz inspired two participatory design session expert wikipedia writer sociologist explained clue used quickly ass trustworthiness article according result propose five metric maturity quality assessment wikipedia article accompanying visualization provide reader important clue editing process glance also report discus result user study conducted two preliminary pilot study show subject trust wikipedia article almost blindly third study show wikipediaviz significantly reduces time required ass quality article maintaining good accuracy goal work support expert process hypothesis generation concerning role gene disease deeper understanding complex interdependency gene important bring gene expression context pathway pathway model biological process available online database database large network decomposed small sub better manageability simplification result loss context pathway interconnected gene occur multiple instance scattered network main goal therefore present relevant information gene expression relation expression pathway multiple pathway simple yet effective way achieve employ two different multiple approach traditional multiple view used large datasets highly interactive visualization technique employed support seamless navigation multiple pathway simultaneously link expression contained gene approach facilitates understanding interconnection pathway enables non relation gene expression data evaluated caleydo group user life science community user asked perform three task pathway exploration gene expression analysis information comparison without visual link conducted four different condition evaluation result show system improve process understanding complex network pathway individual effect gene expression regulation considerably especially quality available contextual information spatial organization rated good presented setup paper present visualization field time multidimensional earthquake accelerograph reading direct volume rendering used depict space relationship seismic reading collected sensor station intuitive way progress seismic wave propagation earthquake event directly observed resulting visualization reveals sequence seismic wave initiation propagation attenuation time energy releasing event provide case study magnitude scale mw chi earthquake taiwan thoroughly recorded earthquake event ever history station recorded event reading event increased global strong record five fold station measured east north vertical component acceleration approximately second sensor network released initial raw data within minute chi mainshock essential visualization system fast data exploring analyzing offering crucial visual analytical information scientist make quick judgment raw data requires preprocessing rendered generated sequence ground wave map regular grid cover entire taiwan island sensor network reading result total ground wave map second interval forming volume data set show visualizing time component data spatially uncovers changing feature hidden data seismic horizon indicate change rock property central geoscience interpretation traditional interpretation system involve time consuming repetitive manual volumetric seeding horizon growing present novel system rapidly interpreting visualizing seismic volumetric data first extract horizon surface preprocessing seismic data interaction user assemble realtime horizon part horizon traditional interpretation system use gradient illumination model rendering seismic volume polygon rendering horizon surface employ realtime gradient forward rendering seismic volume yielding result similar high global illumination use implicit surface representation horizon allowing seamless integration horizon rendering volume rendering present collection novel technique constituting interpretation visualization system highly tailored seismic data interpretation describe visualization process verification cosmological simulation code need code verification stem requirement accurate prediction order interpret observational data confidently compare different simulation algorithm order reliably predict difference simulation result understand dependence input parameter setting verification process consists integration iterative hypothesis comparative feature quantitative visualization validate process verifying time evolution result three different cosmology simulation code purpose verification study accuracy amr method versus n simulation method cosmological simulation range well treatment method destroying tumor similar harmful growth human body utilizes coherence inherently periodic movement affected body part periodic respiratory signal patient objective minimizing damage surrounding normal tissue method require constant monitoring operator observes body motion via projection onto ultrasound imaging plane study synchronism motion respiratory signal keeping attentive eye respiratory signal well ultrasound video entire treatment period often inconvenient burdensome paper propose video visualization technique called cyclestack plot reduces cognitive overhead blending video signal together stack layout visualization reveals inherent synchronism target movement respiratory signal visually highlight significant phase shift either two cyclic phenomenon hope arresting operator attention proposed visualization also provides visual overview post analysis enables educated user quickly effectively skim excessively long process paper demonstrates utility cyclestack plot case study using real ultrasound video addition user study performed evaluate merit limitation proposed method respect conventional way watching video signal side even though motivation proposed visualization improvement medical application use ultrasound core technique discussed potential extended application domain requiring analysis cyclic pattern video paper propose approach clustering data parallel coordinate interactive local operation different many method clustering globally applied whole dataset interactive scheme allows user directly apply attractive repulsive operator region interest taking advantage electricity interaction metaphor clutter reduction cluster detection design enables user interact directly parallel coordinate plot provides great flexibility exploring revealing underlying pattern instant feedback work allows user dynamically adjust clustering parameter reach optimum also supply user graph indicating logical relationship cluster experiment show scheme efficient traditional method performing visual analysis task symbol used scatterplots encode data way appropriate perception human visual channel color believed dominant channel lightness regarded important one three color dimension study lightness perception scatterplots context analytic task requiring symbol discrimination specifically performed experiment measure human performance three visual analytic task outlined circle unframed spot equally sized uniform luminance varied ten eleven equispaced level black white used symbol displayed uniform white background sixteen subject divided two group participated experiment task performance time recorded propose model describe process perception lightness assumed early step complex cognitive process mediate discrimination psychophysical law used describe perceptual mapping different mapping scheme compared regression experimental data result show approximate homogeneity lightness perception exists complex task closely described blended combination two opposite power function assuming either light end dark end lightness scale starting point model yield discriminability scale lightness sparse scatterplots white background paper present novel three process visualize structure point cloud arbitrary dimension get insight structure complexity data set would preferably look e plotting corresponding point cloud unfortunately orthogonal scatter plot work three dimension visualization like parallel coordinate scatterplot matrix also problem handling many dimension visual overlap data entity presented solution tackle problem visualizing point cloud indirectly visualizing topology density distribution benefit approach topology computed arbitrary dimension similar examining scatter plot give important information like number size nesting structure accumulated region view approach alternative cluster visualization create visualization first estimate density function using novel high interpolation scheme second compute function topology mean join tree generate corresponding terrain using topological landscape metaphor introduced weber et al finally augment landscape placing original data point suitable location paper introduce visualization method couple trend chart word cloud illustrate temporal content evolution set document specifically use trend chart encode overall semantic evolution document content time work semantic evolution document collection modeled varied significance document content represented set representative keywords different time point time point also use word cloud depict representative keywords since word word cloud may vary one another time word increased importance use geometry mesh adaptive force model lay word cloud highlight word difference two subsequent word cloud method also ensures semantic coherence spatial stability word cloud time work embodied interactive visual analysis system help user perform text analysis derive insight large collection document preliminary evaluation demonstrates usefulness usability work flow visualization field line often used convey global local structure movement flow one challenge find classify representative field line existing solution follow automatic approach generates field line characterizing flow arranges line single picture work advocate user approach exploring vector field method allows user sketch curve pattern matching field line clustering specifically field line whose view projection similar user drawing identified utilized extract similar field line furthermore employ automatic clustering method generate field template user locate subfields interest semi process leverage user knowledge flow field intuitive user interaction resulting promising alternative existing flow visualization solution sketch interface user effectively dissect flow field make structured visualization analysis presentation new method presented flow visualization based advection reaction model key point method use flow dependent reaction source couple advection reaction dynamic embedded semi advection reaction source provides systematic pattern generation integration along pathlines based idea simple equation stable efficient numerical solution visualization model developed allows visualization lagrangian eulerian structure flexible environment user control system also allows emulation conventional dense texture reaction type visualization technique easily provides solution problem method understanding fluid flow difficult problem increasing importance computational fluid dynamic produce abundance simulation data experimental flow analysis employed technique shadowgraph schlieren imaging century allow empirical observation inhomogeneous flow shadowgraph provide intuitive way looking small change flow dynamic caustic effect schlieren cutoff introduce intensity gradation observing large scale directional change flow combination shading effect provides informative global analysis overall fluid flow computational solution method proven complex recently due fundamental physical interaction light refracting flow field paper introduce novel method simulate refraction light generate synthetic shadowgraph schlieren image time scalar field derived computational fluid dynamic data method computes physically accurate schlieren shadowgraph image interactive rate utilizing combination gpgpu programming acceleration method data probabilistic schlieren cutoff result comparing method previous schlieren approximation presented paper proposes novel visualization approach depict variation different human motion data achieved representing time dimension animation sequence sequential curve locality reference space called motion track representation principal advantage representation standard representation motion capture data generally either keyframed timeline motion map entirety map motion difference along time dimension parallel perceptible spatial dimension time capture primary content source data latent semantic difference difficult visually distinguished clearly displayed favoring effective summary clustering comparison analysis motion database paper introduces new design allow interactive direct manipulation volume data volumetrically rendered image present adaptive volume proxy mesh serf define surface encode geometry physical state volume system performs modeling form direct volume deformation adaptively constructing whole geometric structure background keeping structure transparent user complex high structure locally encoded mesh based user volume interest paper also present scheme volume shading applied combination framework improving reality visualization time deformable object demonstrate example geometry direct volume manipulation application volume exploration surgical simulation paper present volumetric lighting model simulates scattering well shadowing order generate high quality volume rendering approximating light transport inhomogeneous participating medium able come efficient gpu implementation order achieve desired effect interactive frame rate moreover many case frame rate even higher achieved conventional gradient shading evaluate impact proposed illumination model spatial comprehension volumetric object conducted user study participant perform depth perception task result study show depth perception significantly improved comparing illumination model conventional gradient volume shading additionally since volumetric illumination model based gradient calculation also le sensitive noise therefore also applicable imaging modality incorporating higher degree noise instance magnet resonance tomography ultrasound present technique automatically convert small number single volume rendered image data set compact representation data set representation multi image explorable image enables interactive exploration volume data transfer function space without accessing original data achieve automatically extracting layer depicted composited image layer recombined different way simulate opacity change recoloring individual feature result demonstrate explorable image especially useful volume data large interactive exploration take long render due underlying mesh structure desired shading effect original volume data available explorable image offer real image interaction preview mechanism remote visualization visualization large volume data low hardware within mobile device web browser boundary labeling point site uniquely connected label placed boundary enclosing rectangle leader may rectilinear straight line segment result reported literature boundary labeling deal one boundary labeling different site labelled differently certain application boundary labeling however one site may required connected common label case presence crossing among leader often becomes inevitable labeling often high degree confusion visualization paper multi boundary labeling crossing among leader avoided substituting hyperleaders leader applying dummy label copy label minimizing number dummy label becomes critical design issue dummy label required initial setting therefore consider problem minimizing number dummy label multi boundary labeling finding placement label hyperleaders total number dummy label minimized crossing among hyperleaders furthermore number dummy label determined minimizing total hyperleader length well bend hyperleaders also concerned postprocessing procedure paper present polynomial time algorithm one two labeling scheme show correctness theoretical point view addition provide simulated annealing algorithm four labeling scheme objective minimize total number dummy label well total leader length experimental result show four solution look promising compared optimal solution system support web site traffic analysis core business intelligence application many company recent paper remark system especially useful measure user interest relevant concept described web site rather counting user access distinct page forming web site paper extends existing measure conceptual web site traffic analysis describes system called cowa support analysis mean network model graph visualization technology graph drawing algorithmic core user interface cowa force directed heuristic computes simultaneous embedding two non graph heuristic optimizes visualization term crossing resolution user geodesic tendency experimental result case study show effectiveness proposed approach practice information visualization essential making sense large data set often high data visualized collection point space dimensionality reduction technique however traditional method often capture well underlying structural information clustering neighborhood paper describe gmap practical algorithm visualizing relational data geographic map illustrate effectiveness approach example several domain many different approach proposed challenging problem visually analyzing large network clustering one promising paper propose new goal clustering especially tailored hybrid tool namely producing intra graph inter graph suitable highly visualization within different representation convention formalize concept framework class defines desired topological property intra graph x class defines desired topological property inter graph exploiting approach hybrid tool effectively combine different node matrix representation allowing user interactively explore graph expansion cluster without loosing mental map proof concept describe system vhyxy hybrid integrates technique present result case study visual analysis co network present efficient method volume rendering ray casting cpu employ coherent packet traversal implicit bounding volume hierarchy heuristically pruned using preintegrated transfer function exploit empty homogeneous space also detail simd optimization volumetric integration trilinear interpolation gradient lighting resulting system performs well low laptop hardware outperform gpu method order magnitude rendering large volume without level workstation show slower gpu method low volume optimized cpu renderer require lod achieve interactive performance large data set trackball metaphor exploited many application volumetric data need explored although provides intuitive way inspect overall structure object interest inspection tedious cavity occur even impossible therefore propose context navigation technique exploration volumetric data navigation technique polygonal data require information rendered geometry strategy sufficient area volume rendering since rendering parameter e transfer function strong influence visualized structure also affect feature explored compensate effect propose novel image navigation approach volumetric data intuitive use proposed technique allows user perform complex navigation task particular get overview well perform inspection without navigation mode switch technique easily integrated raycasting based volume renderers need extra data structure independent data set well rendering parameter discus underlying concept explain enable navigation interactive frame rate using opencl evaluate usability well performance paper present effective transfer function design multivariate volume providing tightly coupled view parallel coordinate plot md dimension projection plot volume rendered image space design pcp showing data distribution variate dimension md showing reduced dimensional feature integrated seamlessly provide flexible feature classification user without context switching different data presentation proposed interface enables user identify interested cluster assign optical property lasso magic wand tool furthermore sketching directly volume rendered image implemented probe edit feature achieve interactivity octree partitioning gaussian mixture model data reduction technique applied experiment show proposed method effective multidimensional tf design data exploration finding correlation among data one essential task many scientific investigation discovery paper address issue creating static volume classification summarizes correlation connection time multivariate data set practice computing temporal spatial correlation large time multivariate data set prohibitively expensive present sampling approach classifying correlation pattern sampling scheme consists three step selecting important sample volume prioritizing distance computation sample pair approximating volume correlation sample correlation classify sample voxels produce static visualization succinctly summarize connection among correlation volume respect various reference location also investigate error introduced step sampling scheme term classification accuracy domain scientist participated work helped u select sample evaluate result approach generally applicable analysis scientific data correlation study relevant recent advance microscopic imaging technology enabled neuroscientist obtain unprecedentedly clear image neuron extract additional knowledge tangled neuron example connective relationship key understanding information processed transmitted within brain paper introduce recent endeavor neuron navigator integrates neuron image database easy visual interface via flexible user interface nng designed help researcher analyze observe connectivity within neural maze discover possible pathway nng neuron image database researcher perform volumetric search using location neural terminal occupation neuron volume within brain space also presence neuron combination spatial restriction shown well nng result multi collaboration neuroscientist computer scientist nng implemented coordinated brain space drosophila fly brain nng accessible http recent year sophisticated visualization method developed support logical structure time aspect computer clinical treatment plan however visualizing effect applying treatment plan well supporting exploration effect patient condition still largely unresolved task fill gap developed prototype enhances known visualization method communicate process treatment plan application together effect patient condition easily understandable way prototype combine advantage enhanced visual recognition pattern traditional information parameter development thus provides mean ass success failure previously applied treatment plan explore effect applied clinical action patient condition identify sub treatment choice mean help physician optimize treatment choice enable developer clinical practice guideline investigate readjust treatment plan molecular dynamic widely used simulation technique investigate material property structural change external force availability powerful cluster algorithm continues increase spatial temporal extent simulation domain pose particular challenge visualization underlying process might consist million particle thousand time step application domain developed special visual metaphor represent relevant information data set approach typically require detailed domain knowledge might always available applicable propose general technique replaces huge amount simulated particle smaller set representative used visualization instead representative capture characteristic underlying particle density exhibit coherency time introduce loose capacity voronoi diagram generation representative mean gpu parallel algorithm way achieve visualization reflect particle distribution geometric structure original data faithfully evaluate approach using real data set application domain material science thermodynamics dynamical system theory increasing demand world energy consumption oil depletion large reservoir resulted need exploring smaller complex oil reservoir planning reservoir valorization usually start creating model subsurface structure including seismic fault horizon however seismic interpretation horizon tracing difficult error task often resulting hour work needing manually repeated paper propose novel interactive workflow horizon interpretation based well position include additional geological geophysical data captured actual drilling instead interpreting volume slice propose seismic interpretation based well position introduce combination minimal cost path minimal cost surface tracing extracting horizon little user input processing volume based well position rather slice able create piecewise optimal horizon surface interactive rate integrated system visual analysis platform support multiple linked view fast verification exploration analysis extracted horizon system currently evaluated collaborating domain expert present technique visualize global uncertainty stationary vector field topological approach start existing approach uncertain vector field topology extend space number conceptional technical challenge performance visual representation arise order solve develop acceleration finding sink source distribution distribution use overlap corresponding volume find separating structure saddle part approach introduce uncertain saddle boundary switch connector provide algorithm extract visual representation use multiple direct volume rendering test method number synthetic real data set robust analysis vector field established important tool deriving insight complex system field model many analysis technique rely computing streamlines task often hampered numerical instability approach ignore resulting error lead inconsistency may produce unreliable visualization ultimately prevent analysis propose new representation vector field surface replaces numerical integration triangle linear map defined boundary representation called edge map equivalent computing possible streamlines user defined error threshold spite error streamlines computed using edge map pairwise disjoint furthermore representation store error explicitly thus used produce informative visualization given piecewise interpolated vector field recent result show possible map class triangle permitting concise description flow behavior work describes detail computing edge map provides technique quantify refine edge map error give qualitative visual comparison traditional technique visualization flow field geometric primitive often challenging due occlusion inevitably introduced streamlines paper present novel view algorithm minimize occlusion reveal important flow feature three dimensional flow field analyze region higher importance utilize shannon entropy measure vector complexity entropy field form three dimensional volume extracted input vector field utilize view complexity measure view calculation introduce notion maximal entropy projection framebuffer store maximal entropy value well corresponding depth value given viewpoint information develop view streamline selection algorithm evaluate choose streamlines cause minimum occlusion region higher importance based similar concept also propose viewpoint selection algorithm work hand streamline selection algorithm maximize visibility high complexity region flow field current simulation turbulent flame instrumented particle capture dynamic behavior combustion next engine categorizing set many million particle featured history movement position changing thermo state help understand turbulence mechanism introduce dual method analyze data starting clustering time series curve phase space data visualizing corresponding trajectory cluster physical space cluster time series curve adopt model clustering technique two scheme first stage characteristic shape relative position particularly concerned classifying time series curve second stage within group curve clustering conducted based curve change time work perform model clustering semi manner user domain knowledge integrated intuitive interaction tool steer clustering process dual method used analyze particle data combustion simulation also applied scientific simulation involving particle trajectory analysis work effective analysis visualization time multivariate data crucial understanding complex dynamic variable interaction temporal evolution advance made area mainly query visualization correlation exploration solution technique investigate important aspect causal relationship among variable sought paper present new approach analyzing visualizing time multivariate volumetric particle data set study information flow using information concept transfer entropy employ time plot circular graph show information transfer overview relation among pair variable intuitively illustrate influence relation pair variable visualization modulate color saturation opacity volumetric data set present three different visual representation namely ellipse smoke metaball particle data set demonstrate information approach present finding three time multivariate data set produced scientific simulation co collaborative task allow team leverage skill individual member numerous guideline exist develop visualization individual working desktop little known group individual interpret comprehend diverse type visual construct larger display study whether group size impact collective understanding relationship three spatial structure using different type presentation carried three experiment compared individual performance structure understanding task performance group containing two four member consider two alternate visualization technique extracting structure information view animated rotation combination one static plus three static two projection view general study suggest group size increase accuracy cost efficiency result also suggest beyond threshold limit group size performance certain task begin degrade regardless group size participant performed better display presented animation condition instead multiple static view except large group needed relate visualization physical counterpart summarize result term steiner model explaining effect group size task characteristic group performance information exploration process often conducted team expert family member colleague team retrieve information different source verify finally compare discus finding find consensus today support collaborative process limited user often end sharing either single pc one user taking control using separate workstation support tight collaboration limited paper present collaborative information linking visually connects information across private shared application window bridge knowledge gap user present technical infrastructure multi interaction personalized meta large multi display demonstrate personalized visual link connect information across existing application modified minimally invasive manner observational experiment showed information linking help individual deal large display space team switch individual information retrieval joint verification discussion netspeak help writer choosing word writing text check commonness phrase allows retrieval alternative mean wildcard query support query implement scalable retrieval engine return high result within millisecond using probabilistic retrieval strategy result displayed wordgraph visualization textual list graphical interface provides effective mean interactive exploration search result using filter technique query expansion navigation observation indicate three investigated retrieval task textual interface sufficient phrase verification task wherein view support context word choice wordgraph best support exploration phrase context underlying corpus preferred view context word choice seems depend query complexity number wildcards query text stream demand effective interactive method explore dynamic massive data set meanwhile extract valuable information visual analysis paper propose interactive visualization system enables user explore streaming text document without prior knowledge data system constantly incorporate incoming document continuous source existing visualization context minimizing potential energy defined similarity document unlike existing method system us dynamic keyword vector incorporate newly keywords data stream furthermore propose special keyword importance make possible user adjust similarity hence achieve preferred visual effect accordance varying interest also help identify hot spot outlier optimize system performance similarity grid parallel implementation graphic hardware achieves instantaneous animated visualization even large data collection moreover system implement powerful user interface enabling various user interaction data analysis experiment case study presented illustrate dynamic system text stream exploration many data set exist contain geospatial temporal element addition core data requires analysis within data set difficult determine data changed spatial temporal range design study present system dynamically exploring geo change data gtdiff provides visual approach representing difference data within user spatial temporal limit illustrating increase decrease occurred system make extensive use spatial temporal filtering binning geo colour encoding multiple coordinated view highly interactive supporting knowledge discovery exploration analysis data case study presented illustrating benefit using gtdiff analyze change catch data cod fishery coast newfoundland canada present method interactively explore multiple attribute trajectory data using density map image show aggregate overview massive amount data far density map mainly used visualize single attribute density map created two procedure first smoothed trajectory aggregated density field density field visualized approach user explore attribute along trajectory calculating density field multiple subset data density field either combined new density field first visualized combined using widget called distribution map user interactively define subset effective intuitive way supported high graphic hardware user get fast feedback computationally expensive density field calculation show versatility method use case maritime domain distinguish period temporal aggregation find anomalously behaving vessel solve ambiguity density map via drill data risk assessment given generic framework lack domain assumption expect concept applicable trajectory domain well map offer familiar way present geographic data country additional information geology displayed help contour heat overlay paper consider visualizing large dynamic relational data taking advantage geographic map metaphor describe system visualizes user traffic internet radio station last address challenge mental map preservation well issue animated map visualization paper present interactive visual analytics system triple perspective visual trajectory analytics exploring analyzing complex traffic trajectory data user equipped carefully designed interface inspect data interactively three perspective temporal multi view previous work visualization transportation research focused macro aspect traffic flow develop visualization method investigate analyze microscopic traffic pattern abnormal behavior spatial view system traffic trajectory various presentation style directly interactive user brushing together convenient pattern exploration selection ring slider improved themeriver embedded glyph indicating directional information multiple scatterplots time horizontal ax illustrate temporal information traffic flow system also harness power parallel coordinate visualize multi aspect traffic trajectory data three view component linked closely interactively provide access multiple perspective user experiment show system capable effectively finding regular abnormal traffic flow pattern paper discus extension integration statistical concept kernel density estimation scatterplot visualization dynamic data interactive rate present line kernel representing streaming data discus concept kde adapted enable continuous representation distribution dependent variable domain propose automatically adapt kernel bandwith kde viewport setting interactive visualization environment allows zooming panning also present gpu realization kde lead interactive frame rate even comparably large datasets finally demonstrate usefulness approach context three application scenario one studying streaming ship traffic data another one oil gas domain process data operation oil rig streaming operational center third one studying commercial air traffic u spanning dynamic network visualization challenging topic due complexity introduced extra time dimension existing solution problem usually good overview presentation interactive analysis propose paper new approach considers dynamic network central focus node dynamic ego network navigation entire network achieved switching focus node user interaction approach complexity compressed dynamic network greatly reduced without sacrificing network time affinity central focus node result able present dynamic ego network single static view well supporting user analysis temporal network pattern describe general framework including network data pre network trend visualization design layout algorithm well several customized interaction addition show approach also extended visualize event multimodal dynamic network finally demonstrate two practical case study effectiveness solution support visual evidence pattern discovery graph often used encapsulate relationship object node diagram commonly used visualize graph suffer visual clutter large graph edge bundling effective technique alleviating clutter revealing high edge pattern previous method general graph layout either require control mesh guide bundling process introduce high variation curvature along bundle force compatibility calculation scalable propose multilevel agglomerative edge bundling method based principled approach minimizing ink needed represent edge additional constraint curvature resulting spline proposed method much faster previous one able bundle hundred thousand edge second one million edge minute present result study comparing five directed representation use screen node diagram goal work extend understanding tradeoff best practice representation edge directed graph help practitioner choose among different option work applies graph directed link depicted using line connecting node tested five different edge representation chosen carefully based user feedback thoroughly cover directed design space also investigated use pattern compression affect performance subjective user preference article present detailed result regarding significant performance preference difference directed representation provides practical recommendation use present new system visfan visual analysis financial activity network support analyst effective tool discover financial crime like money laundering fraud compared existing system methodology analysis criminal network visfan present following main novelty combine bottom top interaction paradigm visual exploration complex network make possible mix automatic manual clustering allows analyst interactively customize dimension cluster region apply different geometric constraint layout visfan also implement several tool social network analysis clustering example computes several index measure centrality actor network propose notion composite visualization view theoretical model unifies existing coordinated multiple view paradigm strategy combining visual representation geometrical space identify five strategy called cvv design pattern based extensive review literature composite visualization go show design pattern expressed term design space describing correlation two visualization term spatial mapping well data relationship item visualization also discus use design space suggest potential direction future research demonstrate application topological analysis technique rather unexpected domain software visualization collect memory reference trace running program recasting linear flow trace record high point cloud metric space use topological persistence automatically detect significant circular structure point cloud represent recurrent cyclical runtime program behavior visualize recurrence using radial plot display time evolution offering multi visual insight detecting potential candidate memory performance optimization present several case study demonstrate key insight obtained using technique navigation high data space remains challenging making multivariate data exploration difficult effective appealing mainstream application navigation use paradigm metaphor user already familiar one intuitive navigation paradigm interactive route planning connected network employed interface paired prominent high visualization paradigm showing n data undistorted raw form parallel coordinate network interface dimension form node connected network edge representing strength association dimension user interactively specifies node visit system computes optimal route edited manipulated interface route captured parallel coordinate data display dimension ordering configured specified route framework serf data exploration environment interactive presentation platform demonstrate explain justify identified relationship others demonstrate interface within business scenario application progressive refinement methodology make possible elegantly integrate scalable data compression access presentation one approach specifically paper concern effective use progressive parallel coordinate utilized routinely high data visualization discus power typical stage progressive data visualization also utilized fully ppcs different implementation underlying method potential application domain described paper also present empirical result concerning benefit ppc regard efficient data management improved presentation indicating proposed approach able close gap data handling visualization describe practical approach visualizing multiple relationship defined dataset using geographic map metaphor cluster node form country neighboring country correspond nearby cluster aim provide visualization allows u compare two map evolving dynamic process obtained using different relationship case considering multiple relationship e different similarity metric also provide interactive tool visually explore effect combining two relationship method ensures good readability mental map preservation based dynamic node placement node stability dynamic clustering cluster stability dynamic coloring color stability analyzing message stream social blogging service twitter challenging task vast number document produced daily time availability geolocated realtime manually created status update invaluable data source situational awareness scenario work present approach allows interactive analysis location microblog message realtime mean scalable aggregation geolocated text visualization purpose use novel cluster analysis approach distinguish local event report global medium reaction detect spatiotemporal anomaly automatically workbench allows scalable visual examination analysis message featuring perspective semantic layer world map representation novel technique used analyst classify presented event candidate examine global scale formation cosmic structure result action gravity matter expanding universe evolution proceeds velocity field change single almost everywhere space multi complex web region associated formation large structure halo clump filament sheet recently structure investigated primarily via mass density field application paper apply data analysis visualization technique cosmological simulation aim studying multistreaming region using velocity probe compared current practice using density information morphology estimator locating overdense region halo finder show velocity method provide useful supporting well complementary information density field multistreaming correlated contain information new interesting information property large structure may extracted e capturing dynamical behavior possible density estimator incorporating novel method setting threshold velocity estimator study relationship density field represented compact overdense halo different property multistreaming region represented different velocity estimator social medium blog microblogs enable user easily rapidly publish information personal activity interest considered provide valuable information viewpoint sociology linguistics marketing paper proposes novel system analyzing temporal change activity interest blogger visualization phrase dependency structure sentence first extract event represent blogger activity interest analyzing phrase dependency sentence blog archive roughly categorize retrieved event according thematic role experiencer agent location noun within event store dependency database retrieve event involve given topic second present visualization framework exploring temporal change event related topic framework enables user find event topic appear within specific timing drill detail event also enables user compare event different timing multiple topic moreover allows observe overview temporal change set event long change frequency event assist user finding trend implement proposed system five blog archive focused japanese report usefulness system using various example graph intuitive powerful tool present relationship entity efficiently visualizing graph node diagram great challenge due visual clutter induced edge crossing node overlapping many edge bundling method proposed disclose high edge pattern though previous method successfully reveal coarse graph structure relation pattern individual node level overlooked addition many edge bundling algorithm computationally complex prevent scaling extremely large graph paper propose sideknot efficient node bundling method cluster knot edge sideknot light run faster existing algorithm disclose relation pattern individual node standing graph pattern relation neighbor application graph visualization input edge associated target length dealing length challenge especially large graph stress model often employed situation however traditional full stress model scalable due reliance initial shortest path calculation number fast approximation algorithm proposed work well graph result le satisfactory graph intrinsically high dimension node overlap unnecessarily propose solution called maxent model applies principle maximum entropy cope extra degree freedom describe force stress majorization algorithm solves maxent model numerical result show algorithm scale well provides acceptable layout large non graph also potential application scalable algorithm statistical multidimensional scaling variable distance present numerical optimization method visualizing undirected graph method variant force graph drawing sufficient generality adopt different basic force model including kamada fruchterman method achieve efficiency use l method numerical optimization experimental result show method give good performance combined kamada model dynamic graph drawing preserving mental map ensuring location node change significantly information evolves time considered important property algorithm designer many prior experiment attempted verify principle surprisingly little success experiment used several different algorithmic method variety graph interpretation question real fabricated data different presentation method however none result conclusively demonstrated importance mental map preservation task performance experiment measure efficacy dynamic graph drawing different manner look memorable evolving graph rather easy interpret observed previous study found significant difference term response time error rate preserving mental map preserving mental map good idea principle find may always support performance however qualitative data suggests term user perception preserving mental map make memorability task easier qualitative data also suggests may two feature dynamic graph drawing may assist memorability interesting subgraphs remain visible time interesting pattern node movement former supported preserving mental map latter present result user study task performance streamtube visualization used three vector tensor field visualization study used tensor field sampled full diffusion tensor magnetic resonance imaging dataset independent variable include illumination model illumination opengl local illumination texture without motion without task three spatial analysis task depth task determining two marked tube closer user viewpoint visual task marking endpoint tube contact task analyzing tube penetration result indicate global illumination improve task completion time task measured global illumination reduced error participant answer local openglstyle rendering visual task motion present motion contributed spatial understanding task cost longer task completion time high texture pattern led longer task completion time higher error rate result help design lighting model flow diffusion field visualization identify situation lighting efficient accurate mri vivo flow imaging modality potential significantly enhance diagnostics therapy cardiovascular disease however current analysis method demand much time expert knowledge order apply mri clinic larger clinical study one missing piece method allowing gain quick overview flow data main property present line predicate approach sort precalculated integral line capture complete flow dynamic bundle similar property introduce several streamline pathline predicate allow structure flow according various feature useful blood flow analysis e velocity distribution vortex flow path user combine predicate flexibly create flow structure help gain overview carve special feature current dataset show usefulness approach mean detailed discussion mri datasets healthy pathological aorta recent work shown great potential interactive flow analysis analysis path line choice suitable attribute describing path line however still open question paper address question performing statistical analysis path line attribute space way able balance usage computing power storage necessity loose relevant information demonstrate carefully chosen attribute set improve benefit state art interactive flow analysis result obtained compared previously published work finite time lyapunov exponent become widespread tool analyzing unsteady flow behavior computation several numerical method introduced provide trade performance accuracy order decide method parameter setting suitable particular application evaluation different ftle method necessary propose general benchmark ftle computation consists number time flow field error measure evaluating accuracy numerically computed ftle field requires ground truth available realistic flow data set since field generally described closed form overcome introduce approach create non vector field closed formulation ftle field using introduce set benchmark flow data set resemble relevant geometric aspect lagrangian structure analytic solution ftle based ground truth perform comparative evaluation three standard ftle concept suggest error measure based variance field extracted ridge structure dense flow visualization streamlines among popular method numerous application scientific visualization work synthesize benefit approach one framework propose hybrid technique work present novel method dense flow visualization produce high contrast image streamline approach manageable spatial frequency begin introducing notion streamline thickness function thickness wave source compute resulting image coherent wave interference pattern show problem equivalent finding dominant eigenvector sparse matrix employ power iteration numerical scheme demonstrate efficient parallel implementation conclude discussing possible quality improvement strategy extension method domain including flow image line integral convolution popular approach visualizing surface vector field due simplicity high efficiency avoid inconsistency color blur user interaction image approach method use surface parameterization volume texture effect smooth transition often require expensive computational memory cost furthermore method cannot achieve consistent lic result granularity color distribution different scale paper introduces novel image lic surface flow preserve texture coherence user interaction make noise texture different viewpoint coherent propose simple texture mapping technique local robust effective meanwhile approach pre sequence mipmap noise texture coarse manner leading consistent transition model zoomed prior perform lic image space mipmap noise texture mapped onto triangle randomly assigned texture coordinate standard image lic based projected vector field performed generate flow texture proposed approach simple suitable gpu acceleration implementation demonstrates consistent highly efficient lic visualization variety datasets present file layout algorithm flow field improve runtime efficiency streamline computation increasing discrepancy speed processor storage device cost becomes major bottleneck computation reduce cost loading data better spatial locality proved effective also known sequential file access efficient facilitate efficient streamline computation propose reorganize data block file following data access pattern efficient effective prefetching accomplished achieve goal divide domain small spatial block order block linear layout based underlying flow direction ordering done using weighted directed graph model formulated linear graph arrangement problem goal arrange file way consistent data access pattern streamline computation allows u prefetch contiguous segment data time disk minimize memory cache miss rate use recursive partitioning method approximate optimal layout experimental result show resulting file layout reduces cost hence enables efficient streamline computation define apap surface surface perpendicular possible steady vector field present method construct discrete representation since general perfectly perpendicular surface vector field exist propose minimize error metric enforce perpendicularity much possible algorithm construct apap surface deforming seed surface anchored domain point discrete setting minimization result iteratively solving linear least problem integrating locally scaled version vector field definition error metric numerical minimization guarantee minimum zero attained perfectly perpendicular surface exists otherwise minimization converges local minimum independent seed configuration resulting surface least sense perpendicular possible flow apply apap surface interactive flow visualization tool demonstrate number synthetic real flow data set peak finding provides accurate classification direct volume rendering sampling directly local maximum transfer function allowing better reproduction high feature however peak finding technique extend higherdimensional classification work develop new method peak finding multidimensional transfer function look peak along image ray use piecewise approximation dynamically sample transfer function space world sample unidimensional peak finding approach useful specifying transfer function greater precision accurately rendering noisy volume data lower sampling rate multidimensional peak finding produce comparable image quality order better performance reproduce feature omitted entirely standard classification precomputation storage requirement attractive alternative preintegration multidimensional transfer function paper method interactive direct volume rendering proposed computes ambient occlusion effect visualization combine volumetric geometric primitive specifically tube shaped geometric object representing streamlines magnetic field line dti fiber tract proposed algorithm extends recently proposed directional occlusion shading model allow rendering geometric shape combination context providing volume considering mutual occlusion structure represented volume geometry paper propose novel volume illustration technique inspired interference microscopy successfully used biological medical material science decade approach simulates optical phenomenon interference microscopy account light interference transparent specimen order generate contrast enhanced illustrative volume visualization result specifically propose pcvr contrast volume rendering dicvr interference contrast volume rendering corresponding phase microscopy differential interference contrast microscopy respectively without complex transfer function design proposed method enhance image contrast structure detail according subtle change optical path difference illustrate thickness change occluded structure interferometry metaphor addition also develop user interface enable slicing specimen section volume data focus context lens also included system convenient data navigation exploration proposed method based upon widely applied microscopy technique intuitive domain expert explore analyze volume data proposed method feedback domain user suggest proposed technique useful volume visualization approach complimentary traditional one artistic illustration important structure fluid flow long tradition appreciated clearly perceivable instructive still conveying relevant information viewer one important illustrative technique visualization cutaway currently cutaway placed manually using view based approach propose optimize visibility important target feature based degree function doi specified interactive visual analysis e brushing scatterplots show problem placing cutaway box optimally np number box overcome obstacle present intelligent method compute cutaway geometric cutaway object positioned using view objective function optimizes visibility feature order approximate optimal solution use monte carlo method exploit temporal coherence dynamic scene performance part implemented gpu proposed method integrated easily existing rendering framework general enough able optimize parameter besides cutaway well evaluate performance algorithm provide case study vorticity visualization turbulent flow paper propose new accurate technique uncertainty analysis uncertainty visualization based fiber orientation distribution function glyph associated high angular resolution diffusion imaging visualization applies volume rendering technique ensemble odf glyph call sip function diffusion shape capture variability due underlying uncertainty rendering elucidates complex heteroscedastic structural variation shape furthermore quantify extent variation measuring fraction volume shape consistent across noise level certain volume ratio uncertainty analysis visualization framework applied synthetic data well hardi human data study impact various image acquisition parameter background noise level diffusion shape image space method processing method applied volumetric data projected rendered image space filtering tone mapping compositing application domain volume visualization image space method carried efficiently counterpart importantly image space method used enhance volume visualization quality applied together volume rendering method paper present discus application series image space method enhancement confocal microscopy visualization including tone mapping compositing color mapping method easily integrated existing confocal visualization tool fluorender outcome full visualization system meet neurobiologist demand qualitative analysis confocal microscopy data many different application field particle simulation like molecular dynamic used study material property behavior nowadays simulation data set consist million particle thousand time step challenging interactive visualization direct glyph representation particle data important visual analysis process rendering method optimized able work sufficiently fast huge data set however perception implicit spatial structure formed data often hindered aliasing visual clutter especially depth structure grasped better visual cue applied even interactive representation hence present method apply object ambient occlusion based local neighborhood information large timedependent particle data set without need precomputations based density information collected real glyph representation data set visually enhanced without significant impact rendering performance allowing visualize multi particle data set interactively commodity workstation molecular surface established tool analyze study evolution interaction molecule one advanced representation molecular surface called solvent excluded surface present novel simple method representing solvent excluded surface method requires precomputation therefore allows u vary s parameter outright utilize theory implicit surface csg operation compose implicit function representing molecular surface locally function return minimal distance s representation additionally negative value implicit function determine point lie outside s whereas positive one point lie inside describe build implicit function composed three type patch constituting s representation finally propose method visualize iso implicit function mean ray set rendering parameter affecting overall performance material interface free surface topic increasing interest field computational fluid dynamic part reconstructed interface multi simulation behave like classic integral surface known visualization community region surface undergo topological change behave orthogonally expected underlying flow field thus analysis flow field connection material interface shape topology challenging task develop technique facilitates visualization analysis complex material interface behavior time matter track surface parametrization time material interface identify location interaction material interface fluid particle splatting surface visualization technique produce intuitive representation derived interface stability result demonstrate interaction flow field material interface highlighted appropriate extraction visualization technique developed technique aid analysis mixing material interface consistency visualization vascular structure common frequently performed task field medical imaging exist well established applicable method maximum intensity projection curved planar reformation however calcified vessel wall investigated occlusion hinders exploration vessel interior mip contrast cpr offer possibility visualize vessel lumen cutting single vessel along centerline extending idea cpr propose novel technique called centerline reformation capable visualizing lumen spatially arbitrarily oriented vessel necessarily connected tree structure order visually emphasize depth overlap occlusion halo optionally envelope vessel lumen required vessel centerline obtained volumetric data performing scale based feature extraction present application proposed technique focus context setup demonstrate facilitates investigation dense vascular structure particularly cervical vessel vessel data featuring peripheral arterial occlusive disease pulmonary embolism finally feedback domain expert given present new circular cartogram model country drawn polygon circular arc instead straight segment given political map value associated country map cartogram distorted map area country proportional corresponding value circular cartogram model straight segment replaced circular arc order modify area polygon corner polygon remain fixed country circular cartograms aesthetically pleasing appearance cloud snowflake depending whether edge bent outwards inwards make easy determine whether country grown shrunk overall shape show determining whether given map given area realized circular cartogram np problem next describe heuristic method constructing circular cartograms us max computation dual graph map along computation straight skeleton underlying polygonal decomposition method implemented produce cartograms yet perfectly accurate achieve many desired area real example present novel approach text visualization called descriptive non rendering exploit inherent spatial abstract dimension text document integrate non rendering information visualization visualization encodes text data onto model emphasizing relative significance word text physical real relationship word analytic exploration supported collection interactive widget direct multitouch interaction model applied method analyze collection vehicle complaint report national highway traffic safety administration qualitative study demonstrate system support task comparing reliability different model finding interesting fact revealing possible causal relation car part composing hand urban map common design problem avoid overlap geographic feature road building displacing consistently map domain nonetheless automating map design process still challenging task maximally retain depth perception inherent pair parallel line embedded original layout geographic feature paper present novel approach disoccluding important geographic feature creating urban map enhancing visual readability accomplished formulating design criterion constrained optimization problem based linear programming approach mathematical formulation allows u systematically eliminate occlusion landmark road building control degree local map deformation devising objective function minimized various design example together user study presented demonstrate robustness feasibility proposed approach associating region geographic subdivision cell grid basic operation used various type map like spatially ordered treemaps od map case regular shape grid cell allows easy representation extra information region main challenge find association allows user find region grid quickly call representation set region grid grid map network present many field finance sociology transportation often network dynamic structural well temporal aspect present technique extends massive sequence view analysis temporal structural aspect dynamic network using feature data well visualization based gestalt principle closure proximity similarity developed node reordering strategy msv make feature stand enables user find temporal property trend counter trend periodicity temporal shift anomaly network well structural property community star show effectiveness reordering method synthetic real transaction data set dynamic graph increasingly pervasive modern information system however understanding graph change time difficult present two technique simplified visualization dynamic graph using edge bundle first technique us recent image graph bundling method create smoothly changing bundle streaming graph second technique incorporates additional edge data thereby suited visualize discrete graph sequence illustrate method example real large dynamic graph datasets large scale semantic graph social network knowledge graph contain rich useful information however due combined challenge scale density heterogeneity impractical user answer many interesting question visual inspection alone even semantically simple question extended friend also fan favorite band fact require information non number node answer paper propose method combine graph browsing query overcome limitation visual inspection using query main way information discovery graph exploration expand query enables user probe beyond visible part graph bring interesting node leaving view clutter implemented prototype called graphcharter demonstrated effectiveness usability case study user study freebase knowledge graph million node edge graph used represent variety information molecular structure biological pathway computational workflow growing volume data represented graph problem understanding analyzing variation collection graph increasing importance present algorithm compute single summary graph efficiently encodes entire collection graph finding merging similar node edge instead merging node edge exactly use domain comparison function collapse similar node edge allows u generate compact representation collection addition developed method allow user interactively control display summary graph interaction include ability highlight individual graph summary control succinctness summary explicitly define specific node merged show approach generating interacting graph summary lead better understanding graph collection allowing user easily identify common substructure key difference graph paper propose novel volume visualization system enabling local transfer function specification direct painting sketching rendered image wysiwyg style localized transfer function defined scalar topology region specified user intelligent fast feature inference algorithm developed convert user input region specification achieve desirable feature style local transfer function system user manipulate color appearance object volume also apply style transfer generate various illustration style unified input gesture without manual transfer function editing without parameter specification system capable generating informative illustration intuitively highlight user specified local feature multivariate volumetric datasets important science medicine propose transfer function design approach based user selected sample spatial domain make multivariate volumetric data visualization accessible domain user specifically user start visualization probing feature interest slice data value instantly queried user selection queried sample value used automatically robustly generate high dimensional transfer function via kernel density estimation alternatively gaussian tfs automatically generated dimensionality reduced space using sample extracted feature rendered volume rendering view user refine feature using segmentation brush interactivity achieved system different view tightly linked use case show system successfully applied simulation complicated seismic data set paper present user study use depth field depth perception direct volume rendering direct volume rendering phong shading perspective projection used baseline depth field added see impact correct perception ordinal depth accuracy response time used metric evaluate usefulness depth field onsite user study two part static dynamic eye tracking used monitor gaze subject result see though depth field act proper depth cue condition used reinforce perception feature front best result accuracy fast response time correct perception ordinal depth occurs front feature two feature user choose focus perspective projection used volumetric datasets continue grow size continued demand interactive analysis datasets storage device throughput increasing quickly interactive analysis workflow becoming working set ideal workflow working set complexity interactive analysis portion workflow depend primarily size analysis result produced rather size data analyzed past work online analytical processing visualization addressed problem within application context generalized solution wider variety visualization application propose general framework reducing working set complexity interactive portion visualization workflow built top distribution range query well technique within framework able support multiple visualization application transformation applied preprocessing phase workflow enable fast approximate volumetric distribution range query low working set complexity interactive application algorithm adapted make use distribution range query enabling efficient interactive workflow large data show proposed technique enables application scaled primarily term application result dataset size rather input data size enabling increased interactivity scalability deep brain stimulation surgical intervention known reduce eliminate symptom common movement disorder parkinson disease dystonia tremor intervention surgeon place electrode inside patient brain stimulate specific region since region span couple millimeter electrode misplacement severe consequence reliable accurate navigation great importance usually surgeon relies fused ct mri data set well direct feedback patient recently microelectrode recording support navigation measuring electric field patient brain also used propose visualization system fuse different modality imaging data mer patient check well related uncertainty intuitive way present placement information consistent view goal supporting surgeon final placement stimulating electrode describe design consideration system technical realization present outcome proposed system provide evaluation brain activity data often collected use electroencephalography data acquisition modality electric field generated neuron measured scalp although technology capable measuring activity group neuron recent effort provide evidence small neuronal collection communicate distant assembly brain cortex collaborative neural assembly often found examining eeg record find shared activity pattern paper present system focus extracting visualizing potential neural activity pattern directly eeg data using system neuroscientist may investigate spectral dynamic signal generated individual electrode group sensor additionally user may interactively generate query processed reveal area brain may exhibit common activation pattern across time frequency utility system highlighted case study used analyze eeg data collected working memory experiment developed hybrid wind tunnel measurement simulation utilizes actual data acquired real air flow behind square cylinder play important role improving accuracy numerical analysis wind tunnel requires accompanying visual analysis tool effectively peer relationship actual simulated flow field paper attempt exploit augmented reality display end basic idea superimpose computationally mi simulated pressure field onto actual flow velocity structure physically oil misttraced streaklines instantaneously spatial registration two visual source rather straight since fixed cylinder wind tunnel easily identified whereas visualizing mi simulated pressure field characterized sophisticated scheme based differential topology considering fact vortex center located local minimum pressure field minimum surrounded derived topological feature called ridge cycle colorize field adaptively keep track karman vortex street robustly regardless drastic change reynolds number flow field piecewise linear interface calculation one widely employed reconstruction scheme simulation multiphase flow visualization paper focus reconstruction simulation point view present framework analysis reconstruction scheme together implication overall simulation interpreting plic reconstruction isosurface extraction problem first taylor approximation underlying volume fluid field obtain framework error analysis geometric representation reconstruction including flux involved simulation time generalizes plic higher approximation exemplify utility versatility visualization approach several multiphase cfd example last year illustrative method found way flow visualization since communicate difficult information comprehensible way great benefit especially domain audience necessarily flow expertise one domain medical field development mr imaging blood flow measurement lead increased demand easy flow analysis technique goal challenge transfer data simple visualization supporting physician flow interpretation decision making work take one step towards goal present approach illustrative visualization steady flow feature occurring mri data heart aorta like shown manually created illustration restrict visualization main data characteristic depict every flow detail input method flow feature extracted dataset complete set streamlines help line predicate create abstract depiction line bundle selecting set bundle representative reflecting important flow aspect line rendered three arrow fused area represent flow since vortex another important flow information physician identify region mri data display unobtrusive tube structure hatching texture provides visual effect rotational blood flow applying illustration technique diverse flow structure several mri datasets demonstrate abstract visualization useful gain easier insight data feedback medical expert confirmed usefulness revealed limitation work image restricted essential flow feature therefore clearer le cluttered method great potential offer many possible application e comparative visualization also beyond medical domain significant advance made time data analysis visualization mainly improving ability identify temporal trend classify underlying data however ability perform cost data querying indexing often incorporated post serious limitation size time data continue grow paper present new approach unifies data compacting indexing classification single framework achieve transforming time curve representation time data set hierarchical symbolic representation build indexable version data hierarchy create itree visual representation time data hyperbolic layout algorithm employed draw itree large number node provide focus visualization interaction achieve effective querying searching tracking time data set enabling multiple coordinated view consisting itree symbolic view spatial view selecting meaningful feature central analysis scientific data today multivariate scientific datasets often large complex making difficult define general feature interest significant scientific application address problem propose three general spatiotemporal metric quantify significant property data feature continuity co named collectively implemented interactive visualization system investigate complex multivariate time data satellite remote sensing great spatial resolution well real continental power grid monitoring great temporal resolution system integrates metric elegant multi user interaction tool provide various form quantitative user feedback system support iterative user analysis process finding demonstrate metric useful simplifying problem space revealing potential unknown possibility scientific discovery assisting user effectively select significant feature group feature visualization analysis user comprehend problem better design future study using newly discovered scientific hypothesis decision maker often need take account multiple conflicting objective selecting solution problem result potentially large number candidate solution considered visualizing pareto frontier optimal set solution multi problem considered difficult task problem hand span three objective function introduce novel visual approach facilitate coping multi problem propose characterization pareto frontier data task decision maker face reach decision following comprehensive analysis design alternative show semantically self map utilized meet identified task argue newly proposed design provides consistent orientation mapping well appropriate visual representation individual solution demonstrate applicability two real multi case study conclude preliminary empirical evaluation qualitative usefulness assessment contour tree reeb graph firmly embedded scientific visualization analysing univariate field generalize analysis multivariate field data structure called joint contour net quantizes variation multiple variable simultaneously report first algorithm constructing joint contour net demonstrate contour tree individual variable extracted joint contour net investigate visual task solution strategy exploring traditional orthogonal radial node tree layout four orientation non layout well varying difficulty task strategy identified examining eye movement data recorded controlled user study previously conducted burch et al detailed analysis spatio structure pattern eye tracking data employ visual analytics technique adopted related methodology geographic movement data andrienko et al way complement statistical analysis task completion time error rate reported burch et al spatio strategy explain variation completion time identify difference task solution strategy dependent layout type orientation task difficulty furthermore examine difference group participant split according completion time analysis identifies layout took nearly time find task solution node radial layout solution confirmed directly instead frequent cross occurs afterwards main reason impaired performance radial layout agent simulation become key technique modeling simulating dynamic complicated behavior social behavioral science simulation become complex generate increasingly large amount data lacking appropriate tool support become difficult social scientist interpret analyze result simulation paper introduce aggregate temporal graph graph formulation used capture complex relationship discrete simulation state time using formulation assist social scientist identifying critical simulation state examining graph substructure particular define concept gateway inverse terminal capture relationship pivotal state simulation inevitable outcome propose two real computable algorithm identify relationship provide proof correctness complexity analysis empirical run analysis demonstrate use algorithm large social science simulation political power violence present thailand discus broader application atg associated algorithm domain analytic provenance present novel integrated visualization system enables interactive visual analysis ensemble simulation used ocean forecasting simulation sea surface elevation system enables interactive planning placement operation structure illustrate using real simulation gulf mexico structure used oil exploration vulnerable hazard caused strong loop current oil gas industry therefore relies accurate ocean forecasting system planning operation nowadays forecast based multiple spatio simulation resulting multidimensional multivariate multivalued data ensemble data change sea surface elevation good indicator movement loop current eddy visualization approach enables interactive exploration analysis enable analysis spatial domain planning placement structure well detailed exploration temporal evolution chosen position prediction critical ocean state require shutdown rig operation understanding cause protein change shape resulting shape influence function expedite design narrowly focused drug therapy shape alteration often result flexibility change set localized neighborhood may may act concert computational model developed predict flexibility change varying empirical parameter paper tackle significant challenge facing scientist analyzing output computational model namely identify examine compare group interesting neighborhood protein different parameter set difficult task since comparison protein subunit comprise diverse neighborhood often complex characterize simple metric numerous analyze manually present series novel visual analytics approach toward addressing task user scenario illustrate utility approach feedback domain expert confirms effectiveness graph used model relation set object object represented vertex relation edge graph besides vertex relation application domain also relation edge exist new visualization approach support investigation relation type one diagram edge relation visualized curve directly integrated node diagram represents object structure contrast vertex relation illustrated distinguishably edge relation using straight link representation shape link used differentiate relation type weight edge relation mapped width color curve facilitate extensive analysis interrelation approach incorporates several interaction technique used filtering highlighting usability visualization demonstrated two case study application domain bioinformatics financial service readability criterion commonly used measure quality graph visualization paper argue readability criterion necessary sufficient propose new kind criterion generically termed faithfulness evaluating graph layout method propose general model quantifying faithfulness contrast well established readability criterion use example multidimensional scaling edge bundling several visualization metaphor matrix map visualization illustrate faithfulness propose flexgd force algorithm straightline undirected graph drawing algorithm strives draw graph layout encompassing uniform vertex distribution extreme structure abstraction flexible parameterized emphasis put either two drawing criterion parameter determines much edge shorter average distance vertex extending clustering property linlog model flexgd efficient cluster visualization adjustable level energy function flexgd minimized multilevel approach particularly designed work context edge length distribution uniform applying flexgd several real datasets illustrate good quality layout various topology ability algorithm meet addressed drawing criterion response large number existing tree layout generic recently proposed generic approach utilize layout design space pinpoint tree drawing desired characteristic wealth available drawing option parameter design generic layout work well confined set implicit space tree layout far eluded extension explicit node diagram order produce implicit explicit tree layout paper part descriptive nature design space instead take generative approach based operator operator combined operator sequence used different stage layout process small operator set already suffices yield large number different tree layout end present generic tree layout pipeline give example suitable layout operator plug pipeline prototypical implementation pipeline operator presented illustrated space node example furthermore paper present result user study evaluating generative approach realized prototype visual exploration large complex flow field critically important understanding many aero hydro system dominate various physical natural phenomenon world paper introduce flowgraph novel compound graph representation organizes streamline cluster spatial region hierarchically occlusion controllable visual exploration approach work seeding strategy long domain well covered important flow feature captured transforming flow field graph representation enable observation exploration relationship among streamline cluster spatial region interconnection transformed space flowgraph provides visual mapping abstract streamline cluster spatial region various level detail also serf navigation tool guide flow field exploration understanding brushing linking conjunction spatial streamline view demonstrate effectiveness flowgraph several visual exploration comparison task well accomplished using streamline view alone occlusion clutter almost ubiquitous flow flowgraph represents promising direction enhancing ability understand large complex flow field data paper applies poisson method assist interactive exploration steady flow field using data deformation obtain flow flow surface flux optimization surface positioned interactively deformed real according local flow deformed surface particularly useful defining seed structure show gradient computational framework applied obtain parametrizations flow surface way easy define nontrivial seed structure integration flow visualization method additionally flow parametrizations employed view surface lic visualization apply method number data set show effectiveness deformation parametrization seed extraction method interactive flow exploration linear approximation play important role many area employing numerical algorithm particularly field vector field visualization basis widely used technique paper introduce two method extract area two three vector field connected linear flow behavior propose region algorithm extract linear neighborhood certain position region characterized linear flow behavior user approximation threshold first method computes size region given mentioned threshold second method computes quality linear approximation given user n neighborhood scalar field resulting second method therefore called affine linear approximation error isosurfaces field show region close non flow behavior demonstrate expressiveness discus property extracted region using analytical example several datasets domain computational fluid dynamic streamline technique designed based idea property streamlines indicative feature underlying field paper show statistical distribution measurement along trajectory streamline used robust effective descriptor measure similarity streamlines distribution approach present framework interactive exploration vector field streamline query clustering streamline query allow u rapidly identify streamlines share similar geometric feature target streamline streamline clustering allows u group together streamlines similar shape based user selection different cluster different feature different level detail visualized highlight feature flow field demonstrate utility framework simulation data set varying nature size paper proposes vector field visualization mimic sketch representation visualization combine two major perspective large scale trend based strongly simplified field background visualization local visualization highlighting strongly expressed feature exact position component considers vector field spatial derivative derivate asymmetric tensor field allows deduction scalar quantity reflecting distinctive field property like strength rotation shear basis background visualization vector scalar clustering approach local feature defined extremum respective scalar field applying scalar field topology provides profound mathematical basis feature extraction design decision guided goal generating simple read visualization demonstrate effectiveness approach show result three different data set different complexity characteristic hand schematized map traditionally make extensive use curve however automated approach curved schematization previous work focus straight line present new algorithm area curved schematization geographic outline algorithm convert simple polygon schematic crossing representation using circular arc use two basic operation iteratively replace consecutive arc desired complexity reached result restricted arc ending input vertex method steered towards different degree encourage discourage use arc large central angle via single parameter method creates visually pleasing result even low output complexity conducted online user study investigating effectiveness curved schematization compared straight schematization equivalent complexity visual complexity curved shape judged higher using straight line user generally preferred curved schematization observed curve significantly improved ability user match schematized shape moderate complexity unschematized equivalent measuring similarity integral curve fundamental many important flow data analysis visualization task feature detection pattern querying streamline clustering hierarchical exploration paper introduce flowstring novel approach extract shape invariant feature streamlines utilizes string method exploratory streamline analysis visualization solution first resamples streamlines considering local feature scale classify resampled point along streamlines based shape similarity around local neighborhood encode streamline string well shape character construct meaningful word querying retrieval unique feature approach capture intrinsic streamline similarity invariant translation rotation scaling leveraging suffix tree enable efficient search streamline pattern arbitrary length complexity linear size respective pattern design intuitive interface user interaction support flexible querying allowing exact approximate search robust partial streamline similarity matching user perform query either character level word level define character word conveniently customized search demonstrate effectiveness flowstring several flow field data set different size characteristic moving object visualization object attribute commonly represented glyph geographic map area map densely populated object visual clutter occlusion glyph occur propose method solve problem partitioning set object subset visualized using aggregated multivariate glyph show distribution several attribute object heading type velocity choose combination subset glyph design glyph overlap number subset approximately maximal partition maintained updated object move use example maritime domain method applicable wider range dynamic data user study find set representative task method perform significantly worse competitive visualization respect correctness furthermore performs significantly better density comparison task high density data set also find participant user study preference method present flowtour novel framework provides automatic guide exploring internal flow feature algorithm first identifies critical region extract skeleton feature characterization streamline placement create candidate viewpoint based construction simplified mesh enclosing critical region select best viewpoint based viewpoint quality measure finally design tour traverse selected viewpoint smooth efficient manner visual navigation exploration flow field unlike existing work consider external viewpoint unique contribution work also incorporate internal viewpoint enable clear observation lie inside flow field algorithm thus particularly useful exploring hidden occluded flow feature large complex flow field demonstrate algorithm several flow data set perform user study confirm effectiveness approach paper present novel scalable approach visualizing multivariate unsteady flow data lagrangian attribute space projection distance spatial temporal sample evaluated attribute value along advection direction flow field massive sample projected screen space feature identification selection hybrid parallel system tightly integrates mapreduce particle tracer scalable algorithm massive projection designed support large scale analysis result show proposed method system capable visualizing feature unsteady flow couple multivariate analysis vector scalar attribute projection analysis flow data often guided search characteristic structure semantic meaning one way approach question identify structure interest human observer challenge find similar structure datasets different scale orientation paper propose use moment invariant pattern descriptor flow field moment invariant one popular technique description object field image recognition recently also applied identify vector pattern limited directional property flow field contrast previous work follow intuitive approach moment normalization result complete independent set translation rotation scaling invariant flow field descriptor also allow distinguish flow feature different velocity profile apply moment invariant pattern recognition algorithm real world dataset show theoretic result extended discrete function robust way vector field simplification aim reduce complexity flow removing feature order relevance importance reveal prominent behavior obtain compact representation interpretation existing simplification technique based topological skeleton successively remove pair critical point connected separatrix using distance area relevance measure method rely stable extraction topological skeleton difficult due instability numerical integration especially processing highly rotational flow geometric metric consider flow magnitude important physical property flow paper propose novel simplification scheme derived recently introduced topological notion robustness provides complementary view flow structure compared traditional topological approach robustness enables pruning set critical point according quantitative measure stability minimum amount vector field perturbation required remove lead hierarchical simplification scheme encodes flow magnitude perturbation metric novel simplification algorithm based degree theory fewer boundary restriction handle general case finally provide implementation piecewise setting apply synthetic real datasets parallel coordinate often used visualization method multidimensional data set main challenge large data set visual clutter plotting hamper recognition pattern data present edge method using density clustering dimension reduces clutter provides faster overview cluster trend moreover allows rendering clustered line using polygon decreasing rendering time remarkably addition design interaction support multidimensional clustering method user study show improvement classic parallel coordinate plot two user task correlation estimation subset tracing although interactively rotate projected high geometry observe dynamic change traditional visualization method limited highly sensitive choice viewing direction parallel plot supplement visualization scenario providing statistical analysis geometry distinct pair co analysis result visual signature embed geometric structure symmetry thus allows u overview status missing dimension exploring projected geometry paper present blue sampling approach efficient construction continuous parallel plot high geometric surface defined mathematical equation employ parallel plot embedded visual signature assist interactive exploration high geometry typically embedded interactively explore projected geometry observe dynamic change visual signature various geometric property also identified visualized moreover interactively brush plot see counterpart projection assorted geometric property curvature used enhance visual signature scatter plot display relation pair variable given set v variable v pair variable thus number possible pair wise scatter plot therefore even small set variable number scatter plot large scatter plot matrix easily run pixel presenting high data introduce theoretical method testbed assessing whether method used guide interactive exploration high data method based nine characterization distribution orthogonal pair wise projection set point multidimensional euclidean space working directly characterization locate anomaly analysis search similar distribution large splom hundred dimension testbed scagexplorer developed order evaluate feasibility handling huge collection scatter plot wide variety real application generate massive high dimensional categorical datasets datasets contain categorical variable whose value comprise set discrete category visually exploring datasets insight great interest importance however discrete nature often confounds direct application existing multidimensional visualization technique use measure entropy mutual information joint entropy mean harnessing discreteness generate effective visualization conduct user study ass benefit visual knowledge discovery aggregation scalable strategy dealing large network data existing network visualization allowed node aggregated based node attribute network topology advantage however previous system capability enjoy best world paper present oniongraph integrated framework exploratory visual analysis large heterogeneous network oniongraph allows node aggregated based either node attribute topology mixture subset node flexibly split merged hierarchical focus interaction model supporting sophisticated analysis network data node aggregation contain subset node displayed multiple concentric circle onion metaphor indicating many level abstraction contain evaluated oniongraph tool two real case performance experiment demonstrate commodity desktop oniongraph scale million network preserving interactivity analysis static drawing biological pathway still important research tool biologist gerhard michal created seminal drawing metabolic network thus defined canonical representation key pathway kyoto encyclopedia gene genome provides popular static drawing biological network different type used huge number publication drawing widely known immediately recognizable biologist enables collaborative work simplifies communication analysis result automatic layout pathway map complicated fact information available kegg contain entire layout information reference map present fully automated algorithm interactive kegg layout construction algorithm conserve original kegg layout extent possible improving readability removing unnecessary element organism map multiple pathway map laid simultaneously facilitate navigation larger network algorithm support hierarchical layout sub network thus support interactive exploration large datasets drawing highly connected graph difficult read power graph analysis offer alternate way draw graph set node common neighbour shown grouped module edge connected module implies connection member module thus entire graph may represented much le clutter without loss detail recent experimental study shown lossless compression dense graph make easier follow path however computing optimal power graph difficult paper show computing optimal power one module np therefore likely np general case give ilp model power graph computation discus ilp cp technique poorly suited problem instead able find optimal solution much quickly using custom search method also show restrict type search allow limited back provide heuristic better speed better result previously known heuristic describe practical approach visual exploration research paper specifically use title paper dblp database create call map computer science word phrase paper title city map country created based word phrase similarity calculated using co help heatmaps visualize profile particular conference journal base map similarly heatmap profile made individual researcher group department visualization system also make possible change data used generate base map example specific journal conference used generate base map heatmap overlay used show evolution research topic field year individual researcher research group profile visualized using heatmap overlay specific journal conference base map outline modular extensible system term extraction using natural language processing technique show applicability method information retrieval calculation term similarity creation topic map system available mocs research social network analysis shown graph metric degree closeness often used summarize structural change dynamic graph however visual analytics approach proposed help analyst study graph evolution context graph metric paper present novel approach called graphflow visualize dynamic graph contrast previous approach provide user animated visualization graphflow offer static flow visualization summarizes graph metric entire graph evolution time solution support discovery high pattern difficult identify animation individual static representation addition graphflow provides user set interaction create filtered view view allow user investigate particular pattern occurred showcase versatility graphflow using two different datasets describe help user gain insight complex dynamic graph understand recent societal behavior important compare multiple medium affected real event medium affect medium paper proposes novel framework inter comparison visualizing image extracted different type medium extract blog image cluster six blog archive search similar tv shot cluster broadcast news video archive using image similarity visualize image flow timeline space visually interactively explore time sequential change influence among medium resource difference similarity topic become popular blog become popular blog earlier tv new methodology developed discover analyze hidden knowledge massive taxi trajectory data within city approach creatively transforms geographic coordinate latitude longitude street name reflecting contextual semantic information consequently movement taxi studied document consisting taxi traversed street name enables semantic analysis massive taxi data set document corpus hidden theme namely taxi topic identified textual topic modeling technique taxi topic reflect urban mobility pattern trend displayed analyzed visual analytics system system integrates interactive visualization tool including taxi topic map topic route street cloud parallel coordinate visualize probability topical information urban planner administration traveler driver conduct various knowledge discovery task direct semantic visual assist effectiveness approach illustrated case study using large taxi trajectory data set acquired taxi city paper result close collaboration mechanical engineer visualization researcher showcase interdisciplinary work lead new insight progress field case concerned one step product development process goal design mechanical part functional meet required quality measure manufactured standard production method collaboration started unspecific goal first experiment available data visualization method course collaboration many concrete question arose end hypothesis developed discussed evaluated paper facilitate case study validate hypothesis case study consider design reinforcement structure brake lever plastic ribbing three new lever geometry developed basis hypothesis compared reference model validation comprises standard numerical experimental test case new structure outperform reference geometry result promising suggest potential impact product development process also complex scenario advanced composite fiber reinforced polymer promising candidate material future component allow integrating continuously rising demand industry regarding cost function integration weight important structure fiber reinforced polymer individual fiber characteristic strength ductility durability etc large extent determine property final component main contribution paper introduction new system interactive exploration visual analysis fiber property x computed tomography data fiber reinforced polymer presented tool us parallel coordinate define configure initial fiber class using scatter plot matrix linked parallel coordinate initial classification may refined allows analyze hidden relationship individual fiber property view depict resulting fiber classification using polar plot intuitive rendering fiber orientation distribution provided addition two module higher abstraction proposed blob visualization creates hull around fiber similar characteristic fiber metadata visualization allows calculate overlay view containing regional information particular material characteristic proposed system evaluated two group domain expert applying presented concept user feedback show domain expert able efficiently perform task classification fiber visualization fiber length orientation visualization fiber region insight gained forwarded design office well material development simulation order speed development novel composite component gaining insight anatomic co variation help understanding organismic shape variability general particular interest delimiting morphological module generation hypothesis structural co variation undoubtedly highly creative process requires exploratory approach work propose new local anatomic covariance tensor enables interactive visualization explore co variation different level detail stimulating rapid formation evaluation hypothesis effectiveness presented approach demonstrated dataset mouse mandible result literature successfully reproduced providing detailed representation co variation compared state method advent modern smart phone handheld device given analyst decision even general public ability rapidly ingest data translate actionable information paper explore design use mobile visual analytics toolkit public safety data equips law enforcement agency effective situation awareness risk assessment tool system provides user suite interactive tool allow perform analysis detect trend pattern anomaly among criminal traffic civil incident system also provides interactive risk assessment tool allow user identify region potential high risk determine risk user location time system designed iphone environment currently used evaluated consortium law enforcement agency report use system initial feedback crowd sourced crisis mapping relatively new phenomenon platform enables collection visualization real crisis data submitted user social medium tool cellular technology crisis map generally used state non actor sense reference point action current crisis map visualization show location document report short message generated limited representation fails immediately show important content theme document change time result sense becomes time cognitively demanding present new visualization approach geo tag visualization treat tag crowd sourced report spatio textual datasets gtviz embeds map represent spatial property dataset compactly show change key term data source time demonstrate value gtviz two case study controlled user study often necessary analyze spatio data traffic flow air pollution vehicle trajectory city map often used show spatial context various temporal display like time series plot used present change data time paper present novel visualization seamlessly embed temporal display map occlusion visualization spatial temporal attribute data first extend seam carving algorithm broaden road interest map least distortion area embed temporal display road reveal temporal pattern without occlusion map information study various design choice method including encoding time direction temporal display conduct two comprehensive user study validate design decision also demonstrate usability approach case study real traffic flow data major city increasing use gps device location information accessible thus movement people tracked also poi interest information becomes available increasing geo density enable analysis movement behavior present approach enrich trajectory data semantic poi information show additional insight gained using density clustering technique extract frequent destination user movement large e database query available context information foursquare popular location social network enrich destination semantic background gps measurement noisy often one possible destination found movement pattern vary time therefore present highly interactive visualization enable analyst cope inherent geospatial behavioral uncertainty visualization application implicitly run query data retrieve distribution statistical measure derivable distribution distribution based data summary substitute raw data answer statistical query different kind however frequent access raw data longer practical possible answering large number query large data work address issue accelerating range distribution query return distribution axis query region maintaining interactivity query challenging task workload response time query scale data query size paper present framework answering range distribution query arbitrary region near constant time regardless data query size adapt integral histogram based data structure bound workload combination computation communication cost propose two novel transformation data structure decomposition similarity indexing reduce huge storage cost associated addition studying performance range distribution query also demonstrate benefit technique offer visualization application directly indirectly require distribution multidimensional projection technique provide appealing approach multivariate data analysis ability translate high data low representation preserve neighborhood information recent year pushed ever increasing data complexity many area numerous advance technique observed primarily term computational efficiency support interactive application achievement made possible due introduction concept control point used many different multidimensional projection technique however little attention drawn towards process control point selection work propose novel multidimensional projection technique based radial basis function method us rbf create function map data low space interpolating previously calculated position control point also present built method control point selection based least square technique demonstrate proposed selection process allows technique work control point retaining projection quality avoiding redundant control point demand interactively designing image feature space increasing due ongoing need image retrieval recognition labeling although conventional method provide interface locally rearranging feature space category global manipulation still missing thus manually rearranging overall image categorization usually requires time task paper present novel approach exploring image database manipulation bi feature space representation upper lower representation characterize global category local feature image respectively approach upper space describes similarity relationship among underlying category extracted bag model lower space encodes closeness pair image within category key idea behind approach associate relationship two feature space two graph representation project onto screen space using pivot md user manipulation experimental result provided demonstrate approach allows user understand entire structure given image dataset reorganize layout according preference locally globally interactive data exploration manipulation often hindered dataset size data aggravated occlusion important adjacency entangled pattern challenge make visual interaction via common filtering technique hard describe set realtime multi data deformation technique aim help user easily select analyze eliminate spatial pattern technique allow animation view configuration semantic filtering view deformation data subset selected step along animation data filtered deformed reduce occlusion ease complex data selection technique simple learn implement flexible real interactive datasets ten million data point demonstrate technique three domain area image segmentation manipulation medical volume exploration astrophysical exploration paper present algorithm called clue cluster expander compound graph expands cluster node compound graph clue designed work multiple layout algorithm orthogonal layout algorithm sugyima layout algorithm etc keep unexpanded node relative layer slight change original coordination help user maintaining mental map underlying graph applied proposed algorithm navigating compound graph representing failure mechanism embedded system also performed brief user evaluation study order measure user acceptance level regard application clue compound graph result show high acceptance ratio user different background indicates intuitiveness feasibility approach past year research web information visualization focusing present web information end user effectively efficiently nevertheless increasing size information web become difficult visualize information inherent relationship among information web graph messy large end user need allow extra effort mine interested information existing web graph produce relationship among node based structural linkage information result producing similar visualization different end user approach lack effectiveness considering diversity user need work encounter issue personalization visualization web graph propose different web graph capitalizes user interest develop relationship among node web graph example show web graph represent web information end user effective effortless information search display data related graphical object map well studied problem cartography several approach published adapting idea boundary labeling using focus region face problem emphasizing additional information label object within focus region even region might moving propose four different approach support effective solution various way discus advantage drawback algorithm regard dynamic nature focus region place label preserve mental map analyzed drawing demonstrate method applying biological network formal concept analysis derives multiple class hierarchy formal context number class bounded exponential function number object attribute context support interactive analysis large formal context paper exploit divide technique discovers hierarchical structure amenable formal context hierarchical structure used expedite enhance layout user interaction concept lattice principal contribution dual use discovered hierarchical structure scalable interactive fca large network become increasingly challenging manage vital system administrator network manager able analyze vast amount log data order detect suspicious behavior pattern possibly due malicious user faulty device intrusion detection system log provide large number warning exactly alarm true others false importantly underlying cause still difficult know bridge gap network log anomaly discovery design implement visualization tool combine multiple commodity visualization minimum learning curve individual view well understood effect view analyzing network anomaly well studied since visualization technique advantage well limitation addressing particular task show view combined linked together may provide effective lightweight network anomaly analysis tool web open platform may simplify network administration well promote collaborative analysis among researcher dns name system essential component functionality internet convert domain name ip address security dns related whole internet dns query log file provide insight dns security paper propose interactive visual analysis system dns log file intuitively detect anomaly dns query log theme river based ranking visualization linked heat tree map user could easy identify anomaly analyze regional temporal feature help administrator figure reason moreover feature dns query time region could also analysis system transfer function design volume visualization challenging problem due huge design space work present system capable integrating transfer function design result group user specified volume dataset intermediate final transfer function design many user different background collected representation transfer function feature space called transfer function map constructed volume data set md projection collected transfer function sample proposed transfer function map interaction including flexible navigation transfer function feature space transfer function design recommendation developed propose new application stochastic point rendering recently proposed implicit surface large laser point data specifically propose scheme apply rendering transparent fused visualization recent large complex laser data cultural asset scheme us point directly acquired using laser scanner rendering primitive laser data consist point pre stage take minute rendering stage executable interactive frame rate encounter rendering artifact originating indefiniteness depth order rendering primitive fused visualization various visual assistant also possible demonstrate effectiveness scheme visualizing campus building culturally important festival float paper propose technique rendering multiple irregular volume using stochastic approach approach extend stochastic projected tetrahedral algorithm treat irregular volume spt algorithm sorting volume rendering algorithm project tetrahedral cell onto image plane control particle rendering using opacity probability rasterization process rasterization process particle depth determined pre location may front back middle point ray segment spt algorithm determination cause prominent artifact treating multiple volume although cause problem treating single volume artifact occurs spt algorithm control particle stochastically image plane depth direction volume rendering process assumed number particle follows poisson distribution along segment viewing ray intersects tetrahedral cell particle spacing follows exponential distribution case construct cumulative distribution function particle spacing develop technique calculating nearest particle along interval apply technique multiple volume confirm effectiveness lagrangian coherent structure time surface highlight area flow field neighboring advected particle diverge converge detection understanding structure important part many application oceanography need predict dispersion oil material ocean one widely used tool revealing lagrangian coherent structure calculate finite lyapunov exponent whose maximal value appear ridgeline reveal lagrangian coherent structure paper explore alternative formulation lyapunov exponent computing lagrangian coherent structure development supercomputer successfully helped u carry complicated simulation exploded size dataset visualizing kind large dataset reducing data size using compression method one useful approach moreover parallelization compression algorithm greatly improve efficiency resolve limitation memory size however parallel compression algorithm interprocessor communication indispensable also bottleneck problem especially general case number processor power parallel pod orthogonal decomposition compression algorithm example number time step must power binary swap scheme method fully resolve problem low computational cost popular paper proposed approach called combination approach simply implemented also reach high performance parallel computing algorithm furthermore method obtain best balance among parallel computing processor achieved transferring non problem power problem fully use best balance feature binary swap method evaluate approach applying parallel pod compression algorithm k computer system grow increasing pace keep demand larger data ever detailed simulation failure also occur frequently various level traditionally fault tolerance visualization either disregarded due performance reason handled analogously state art high performance computing example rollback paper summarize inherent challenge distributed visualization well applicability difference fault tolerance method used hpc distributed visualization derive complementary method increase resilience system present visualization specific method failure semantics masking eit impedance tomography novel imaging method visualizing impedance change thorax mainly influenced breathing heart beating unfortunately technique poor image resolution improve quality eit image computer scientist medical researcher working together image reconstruction process data background different exist difficulty communication assessing data image provide visualization thorax containing ct tomography eit image closing gap multimodal semiabstract suitable group show rendered thorax volume embedded eit movie eit image reconstruction contains several computational intermediate stage used improve volume rendering transfer function additionally isosurfaces show context structure understand spatial coherence easily visualization designed interactive remove visual clutter provide direct view lung focus interest result show visualization able give basis discussion medical researcher well computer scientist phenomenon discovered long discussion easier detect furthermore possible obtain new knowledge address question term eit providing sustainable urban water distribution system complex process involving decision maker authority well public opinion process successful visualization support analyzing comparing different scenario therefore study several visualization technique combined provide effective representation water distribution system main parameter visualize nodal pressure first nodal pressure value study area obtained function operation time using u environmental protection agency epanet model time step represented small multiple map order improve visibility node node along edge aggregated represented single node circle view representation used visualize two value nodal pressure variable finally water network mapped onto grid allows user see every node individually without overlap result show combination visualization technique illustrate water network scenario beneficial decision maker provide sustainable urban water service present visual analysis application protein simulation allows researcher investigate dynamic data interactively special focus lie analysis cavity binding site play critical role protein function cavity extracted classified surface area cavity diameter pocket channel computed provide evidence accessibility value also plotted graph quantitative time analysis dynamic simulation data set cavity tracked show stability time user provided range application parameter interactively adjust analysis algorithm sequence diagram show structure protein additional annotation like binding site furthermore view support brushing linking consistent selection filtering algorithmic step implemented run interactively commodity workstation result user immediately see effect parameter change enables real analysis running simulation visualization study propose novel integrated visualization system enables interactive visual development habitat suitability index model predicting location fishing ground system enables interactive selection variable highly correlated fish catch recorded fisherman vessel illustrate use system using real simulation pacific ocean vortex structure used fishing ground exploration crucial exploring fishing ground drive nutrient ocean floor sea surface ability locate fishing ground therefore depends accurate ocean forecasting system calculating ocean model variable temperature salinity velocity etc currently forecast based multiple spatio simulation produce multidimensional multivariate result critical point ocean model good indicator likelihood finding microscale vortex visualization approach enables interactive exploration analysis critical point proposed system make possible analyze vortex structure spatial domain explore fishing ground sea surface detail performance attribute energy use natural ventilation becoming rapidly important design modern building basis improvement existing visualization technique application domain provide detailed task requirement analysis using feedback architect state visualization strategy used building performance simulation result evaluated comparing quality aspect derived assessment specifically reveals shortcoming respect applied technique visualizing spatiotemporal data therefore discus potential utilizing visualization technique meet identified prerequisite reveal future direction based finding traffic incident road accident vehicle breakdown major source travel uncertainty delay mechanism cause heavy traffic fully understood traffic management controller tasked routing repair clean crew clear incident often time pressure imperfect information aid decision making help understand past incident affected traffic propose traffic origin simple method visualize impact road incident congestion traffic incident occurs mark incident location expanding circle uncover underlying traffic flow map end circle recedes directs attention upcoming event also allows u observe impact traffic incident vehicle flow immediate vicinity cascading effect multiple incident road network illustrate technique using road incident traffic flow data singapore email archive contain rich information interact different contact event evolve throughout time making sense archived message good way understand thing evolved progressed past although much work devoted email visualization work focused presenting one two aspect email archive discovering evolution email event relationship email owner contact time paper present email map email visualization integrates information event contact single view enabling user make sense email archive complementary contextual information two visualization component designed portray complex information within email archive event flow contact track event flow illustrates evolution past event helping user grasp high picture pattern email archive contact track reveal interaction email owner contact order clearly visualize video structure provide user video content review retrieve system firstly apply total variation metric video shot segmentation propose novel video visualization method paper firstly video data visualization design principle proposed based analysis video data characteristic secondly tv metric proposed measure distance two video frame video shot detected tv metric video key frame extracted k algorithm finally according proposed video visualization design principle spiral video visualization method proposed experimental result show proposed video visualization method express video content whole structure directly clearly work propose visual analytic system analyzing event weibo chinese microblog service build system consists two interface web online visualization interface public user offline expert visual analytic system wrap online one provides additional analysis function online interface provides intuitive powerful retweet tree visualization inspires user creativity expert system adopts public user analysis result collected web interface visualize analyze weibo event deeper extent paper outline non user visualisation mode information visualisation commonly practiced artist designer neuvis wicked problem account design constraint related affective emotional response user contrast neuvis tame complex problem associated scientific visualisation example scientific visualisation neuvis show challenge neuvis overcome collaboration artist scientist current research design lab university sydney neuvis aim map different level cognitive emotional response different mode visualisation high resolution ogcm general circulation model reproduces various type complex structure defined multivariable warm current high density water mass order understand characteristic feature necessary visualize explore multivariable dataset therefore applied visual analytics especially multi transfer function visual feature exploration represent characteristic ocean structure case study show visual analytics effective understanding ocean simulation dataset arterial disease sclerosis aneurysm known related local blood flow indeed decade related study recently aid computational approach patient basis yet translated clinical medicine achieve computational blood flow originates theoretical computational fluid dynamic understanding underlying knowledge may beyond scope medical professional engineer fully aware barrier interface clinical medicine manage translate evolving technology special focus balancing efficacy safety visual analytics facilitate overcome academic gap engineering clinical medicine paper deal recent development visual analytics malignant blood flow cerebral aneurysm medical professional appreciating cultural object e artistic statue traditional vehicle historical building city visualization mainly used following style realistic computer graphic artistic computer graphic interactive virtual reality however academic studying cultural object style insufficient namely need analytical style extract invisible veiled feature visualized target make analytical visualization important recent rapid development range technology digitally archive real object example modern laser scanning produce point data one measurement extract useful information large complex data need precise transparent visualization superposition visual assistant time visualization talk clarify role visualization technique mandatory employ scientific method policy currently many country investing heavily policy process development ict purpose policy cycle many opportunity scientific method effectively applied especially agenda setting phase important visualize cognitive structure citizen focus evaluation grid method figure citizen requirement construct cognitive structure represented graph node arc represented construct relationship respectively develop predictive cognitive model useful agenda setting purpose employ multivariate visualization technique figure meaningful relationship cognitive model author laboratory many joint project analysis visualization temporal daily life data transaction credit card web access log walking path public space measurement building management population temperature electric power often discover interesting behavior visualization result feel several result demonstrate japanese specific behavior mentality talk introduces visual analytics result denote japanese specific behavior mentality due space limitation manuscript introduces two example though talk may introduce example illustrating difference dti fiber model important purpose group comparison atlas construction uncertainty analysis standard approach show difference dti model space either voxel fiber based comparison proposes new method embeds high fiber model continuous map normalized space built fiber targeted fiber model one advantage previous embedding approach method represents fiber model unified coordinate system taking cluster routine ensuring closely located fiber different model projected close point map using projected map subtle difference difficult distinguish space clearly displayed individual fiber model effectively summarized applied method group comparison analysis various fiber model contribution point direction visualization research context particular approach visual analytics approach characterized holistic view data integrated visual representation main requirement multivariate micro layered abstraction direct interaction information control occlusion present major challenge visualizing flow field using streamlines displaying many streamlines creates dense visualization filled occluded structure displaying stream risk losing important feature ideal streamline exploration approach visually manipulate cluttered streamlines pulling visible layer apart revealing hidden structure underneath paper present customized deformation algorithm interactive visualization tool minimize visual cluttering algorithm able maintain overall integrity flow field expose previously hidden structure system support mouse direct interaction manipulate viewing perspective visualize streamlines depth using lens metaphor different shape select transition zone targeted area interactively user move focus examine flow field freely generalize framework moment introduce definition invariant three vector field use method moment normalization shown useful two dimension using invariant moment show search pattern field independent position orientation scale first order vector moment tensor construct complete independent set descriptor test invariant query synthetic real world flow field paper introduce novel flow visualization technique arbitrary surface new technique utilizes closest point embedding represent surface allows accurate particle advection surface well support unsteady flow line integral convolution technique surface global approach faster previous parameterization technique prevents visual artifact associated image approach paper present accurate parallel implementation unsteady flow line integral convolution high visualization large time flow approach differs previous implementation using novel value scattering mechanism parallelize uflic designing pathline reuse strategy reduce computational cost pathline integration exploiting massive parallelism modern graphical processing unit proposed method allows real dense visualization unsteady flow high spatial coherence clustering often first step trying make sense large data set wide family cluster analysis algorithm namely hierarchical clustering algorithm provide partition data set hierarchy cluster organized binary tree known dendrogram dendrogram classical node representation used expert various task like decide subtrees actual cluster cutting dendrogram given depth give cluster name inspecting content etc present dendrogramix hybrid tree interactive visualization dendrograms superimposes relationship individual object hierarchy cluster dendrogramix enables user task involve cluster individual object impracticable classical representation like explain particular object belongs particular cluster elicit understand uncommon pattern object could classified totally different cluster etc sensemaking task supported consistent set interaction technique facilitates exploration large clustering result edge bundling method reduce visual clutter dense occluded graph however existing bundling technique either ignore edge property direction data attribute otherwise computationally scalable make unsuitable task exploration large trajectory datasets present new framework generate bundled graph layout according numerical edge attribute direction timestamps weight propose gpu implementation linear number edge make algorithm applicable large datasets demonstrate method application analysis aircraft trajectory datasets eye trace paper introduce novel visualization method allows people explore compare refine major community large network first detect major community network using data mining community analysis method statistic attribute community relational strength community boundary node connecting community computed stored propose novel method based voronoi treemap encode community polygon relative position polygon encode relational strength different community attribute encoded polygon shape size color corner method introduced adjust smoothness polygon based certain community attribute accommodate boundary node gap polygon widened polygon algorithm boundary node conveniently embedded newly created space method efficient enabling user test different community detection algorithm fine tune result explore fuzzy relation community interactively case study two real data set demonstrate approach provide visual summary major community large network help people better understand characteristic community inspect various relational pattern community classic method depict graph node diagram vertex associated object edge connect related object however node diagram quickly appear cluttered unclear even moderately sized graph position node fixed suitable link routing option reduce clutter present novel link clustering routing algorithm respect desired refines user cluster link cluster defined priori cluster based geometric criterion based well pair decomposition route link cluster individually sparse visibility spanner completely avoid ambiguity draw individual link ensure clustered link follow path routing graph prove cluster induced wspd consist compatible link according common similarity measure formalized holten van wijk greedy sparsification visibility graph allows u easily route around obstacle experimental result visually appealing convey sense abstraction order virtual reality researched many way spatial scientific visualization comparatively little explored visualization abstract kind data particular stereoscopic vr environment graph visualization applied limited extension standard technique using stereoscopy highlighting work explore new immersive approach graph visualization designed specifically virtual reality environment visualizing dynamic graph challenging due difficulty preserving coherent mental map changing graph paper propose novel layout algorithm capable maintaining overall structure sequence graph laplacian constrained distance embedding method work online maintains aesthetic individual graph shape similarity adjacent graph sequence preserving shape graph component across different time step method effectively help user track gain insight graph change two datasets tested demonstrate effectiveness algorithm modern social network often dynamic multi however currently little studied incorporate aspect simultaneously support visual analytic task complex social network present visual analytic framework dynamic multi network prototype implementation called multistory system includes two new visualisation method altercluster interarc designed dynamic network multiple relation system evaluated two case study using social network mit reality common demonstrate effectiveness system support variety visual analytical task dynamic multi network visualization hierarchical data great importance information visualization present variational circular treemaps novel layout algorithm solving disk packing continuous optimization problem variational circular treemaps achieve higher space utilization ratio compared traditional circular treemaps support natural interaction focus distortion drill roll operation data navigation experimental result show effectiveness method visualization interaction understanding origin pattern different region city important urban planning work based taxi gps data propose od novel visual design associated analysis tool explore od pattern user define region taxi trip starting ending region selected grouped od cluster hybrid circular visual design od allows user explore dynamic pattern od cluster including variation traffic flow volume traveling time proposed tool support convenient interaction allows user compare correlate pattern different od cluster use study real data set demonstrates effectiveness proposed od interaction trajectory need analyzed various domain gain insight movement pattern interaction often take place delayed response propose approach analyze visualize delayed response two trajectory recorded simultaneously sampling rate central approach computation matching trajectory delay space also introduce new similarity measure trajectory combine directional spatial characteristic evaluate approach experimentally implemented prototype visual analytics tool applied tool two datasets feature integration theory classic visual attention model provides theoretical foundation optimizing visual communication paper proposes set glyph named roseshape aim improving user visual information search visualization roseshapes integrate rich easy attribute mapping data variable enabling automatic unconscious information processing pre stage case study illustrating level education funding united state validates feasibility approach present electronic radially instantaneous coordinate animation visualization approach represent time evolution electronic structure data nuclear coordinate resulting ab initio molecular dynamic simulation developed erica order enable chemist analyze aimd simulation interaction two hydroxyl radical water consequently illustrate erica using simulation discus erica generalized aimd simulation using erica chemist gained new insight hydroxyl radical chemistry radviz radial visualization technique map data multiple dimensional space onto planar picture dimension placed circumference circle called dimension anchor reordered reveal different pattern dataset extending number dimension enhance flexibility placement da explore meaningful visualization paper describe method rationally extends dimension multiple new dimension radviz method first calculates probability distribution histogram dimension mean shift algorithm applied get center probability density segment histogram dimension extended according number segment histogram also suggest using dunn index find optimal placement da better effect visual clustering could achieved dimension expansion radviz finally demonstrate usability approach visually analysing iris data two datasets text visualization become growing increasingly important subfield information visualization thus getting harder researcher look related work specific task visual metaphor mind paper present interactive visual survey text visualization technique used purpose search related work introduction subfield gaining insight research trend describe taxonomy used categorization text visualization technique compare approach employed several survey finally present result analysis performed entry data increasing amount data generated location based social network twitter flickr foursquare currently drawing attention urban planner new source data contains valuable information behavior inhabitant city making data accessible urban planning domain add value decision making process however analysis spatial temporal characteristic data context urban planning ongoing research problem paper describes ongoing work design development visual exploration tool facilitate task proposed design provides approach towards integration visual exploration tool capability visual query system multilevel perspective multiple spatial scale temporal resolution implicit lbsn data preliminary discussion design potential insight gained exploration analysis data proposed tool presented along conclusion future work continuation work work introduce senticompass exploring comparing sentiment time twitter data visualization design combine psychology model affect emotion time tunnel representation illustrate visualization design two case study conducted demonstrate effectiveness senticompass achieving various task related temporal sentiment affective analysis tweet interactive demo system available http car manufacturing quality management control important part series process series production many part controlled various way stage assembly tactile measurement mostly restricted point inspection plan restriction apply optical measurement propose method based theory pareto set topological analysis cope large amount data produced optical measurement find point interest measured surface addition inspection plan describe method automatically detects area systematic error component visualizes triangulated surface visualization help expert decide whether detected feature severe enough added inspection plan paper describes effort create new visualization exploiting hierarchical scalar topology first build hierarchical topology synchronously constructing simplifying contour tree morse complex scalar field introduce three algorithm based hierarchical topology topology multi contouring overview provided scalar field extracting iso simplified ct tracing approximate contour across m complex cell topology based spaghetti plot uncertainty seeding scheme based hierarchical topology visualizing uncertainty among ensemble scalar data virtual ribbon new scheme visualizing multivariate data invented overlapping visual ribbon encode scalar variation region covered uniform contour compare new approach current alternative time surface versatile tool visualise advection deformation flow field due complex flow behaviour involving stretching shearing folding straightforward mesh representation surface develop artefact degenerate quickly common counter rely refinement adaptive insertion new particle lead unpredictable increase memory requirement propose novel time surface extraction technique keep number required flow particle constant providing high level fidelity enabling straightforward load balancing solution implement particle relaxation procedure make use local surface metric tensor model surface deformation combine accurate bicubic surface representation provide artefact surface visualisation demonstrate evaluate benefit proposed method respect surface accuracy computational efficiency ray casting algorithm major component direct volume rendering exhibit inherent parallelism making suitable graphic processing unit however blindly mapping ray casting algorithm gpu complex parallel architecture result magnitude performance loss paper novel computation mapping strategy called warp marching texture iso volume rendering introduced evaluate compare new strategy commonly used existing mapping strategy texture cache performance load balancing two major evaluation factor since significant consequence overall rendering performance series real data experiment conclude texture cache performance two computation mapping strategy significantly affected viewing direction warp marching performs better balancing workload among thread concurrent hardware component gpu massive open online course platform becoming increasingly popular recent year thousand student watching course video enormous amount clickstream data produced recorded moocs platform course large data provide great opportunity instructor educational analyst gain insight online learning behavior unprecedented scale nevertheless growing scale unique characteristic data also pose special challenge effective data analysis paper introduce vismooc visual analytic system help analyze user learning behavior using video clickstream data mooc platform work closely instructor two coursera course understand data collect task analysis requirement complete user design process employed design develop vismooc includes three main linked view list view show overview clickstream difference among course video content view show temporal variation total number type click action along video timeline dashboard view show various statistical information demographic information temporal information conduct two case study instructor demonstrate usefulness vismooc discus new finding learning behavior urban microclimate research mobile transects utilized observe relationship atmospheric variable urban environment however data set resulting mobile measurement complex spatially dependent multivariate often multitemporal time spatial context observation field view area represented physically complex dynamic property make analysis visualization challenging present prototype visualization framework assist researcher analysis mobile transect measurement system enables user visualize explore observation wall delineate transect route map observed attribute stacked upon within wall ribbon facilitate qualitative analysis spatial variability multivariate correlation relationship observation spatial context interactively explored moving slider along transect route observation track spatially contextual source area displayed linked view fraction land cover class contained within source area qualitative analysis capability complemented interactive clustering interface allows classification transect segment according coherent pattern multivariate relationship user set observation framework developed team comprising visualization urban microclimate researcher case study show utility specialized application rise virtualization cloud network various scale degree complexity new approach managing infrastructure required network relationship among component arbitrary cardinality n making challenging administrator investigate component influence others paper present cloudgazer scalable visualization system allows user monitor optimize cloud network effectively reduce energy consumption increase quality service instead visualizing overall network split graph semantic perspective provide much simpler view network cloudgazer multiple coordinated view system visualizes either static live status information component perspective reintroducing lost inter relationship demand using dynamically created inlay demonstrate effectiveness cloudgazer two usage scenario first based real network domain partner static performance parameter used find optimal design second scenario use vast challenge dataset demonstrate system employed live streaming data social medium monitoring crisis intelligence brink commercialization widespread adoption close gap research application paper present result broad case study visual analytics social medium situation awareness asked twenty domain expert disaster response critical infrastructure management investigate various crisis intelligence task based actual twitter data recorded german flood end scatterblogs visual analytics framework platform provides reference implementation tool technique ongoing research given integrated toolbox paper present review domain expert individual performance system comment usefulness applicability capability result questionnaire participant complete based evaluation answer question visual analytics social medium shape tomorrow crisis intelligence work introduces metatracts novel method extracting visualizing individual fiber bundle weaving pattern x computed tomography scan endless carbon fiber reinforced polymer proposed work flow designed analyze unit cell cfrp material integrating recurring weaving pattern designed handle xct scan low resolution individual fiber visible barely visible first coarse version integral curve used trace subsection individual fiber bundle woven cfrp material call section metatracts second step extracted fiber bundle section clustered using two approach first orientation proximity tool generate volumetric representation well surface model extracted fiber bundle exported analysis evaluate proposed work flow number real world datasets demonstrate metatracts effectively robustly identifies separate different fiber bundle work advance research effort combining casual sketching approach whiteboards machine computing power present sketchinsight system applies familiar collaborative feature whiteboard interface accurate data exploration capability interactive visualization sketchinsight enables data analysis fluid interaction allowing people visually explore data drawing simple chart directly manipulating addition report result qualitative study conducted evaluate user experience exploring data sketchinsight expanding understanding people use pen touch digital whiteboard data exploration also discus challenge building working system support data analytic capability pen touch interaction freeform annotation high label layout critical effective information understanding consumption existing labeling method fail help user quickly gain overview visualized data number label large visual clutter major challenge preventing method applied real application address propose context label layout measure reduce visual clutter layout process method formulates clutter model using four factor confusion visual connection distance intersection based clutter model effective clutter labeling method developed generate clear legible label layout different visualization applied method several type visualization result show promise especially support uncluttered informative label layout spatial temporal resolution time simulation become high possible process store data every time step due high computation storage cost although using uniformly data visualization common practice important information un data lost currently linear interpolation popular method used approximate data stored time step pathline computation however error interpolated velocity time dimension accumulate quickly make trajectory rather unreliable inform scientist error involved visualization important quantify display uncertainty importantly reduce error whenever possible paper present algorithm model temporal interpolation error error reduction scheme improve data accuracy temporally data show possible compute polynomial regression measure interpolation error incrementally one sequential scan time flow field also show empirically data sequence fitted least regression error approximated gaussian distribution end position particle trace stored show error modeling scheme better estimate intermediate particle trajectory stored time step based maximum likelihood method utilizes forward backward particle trace although vortex analysis detection extensively investigated past none existing technique able provide fully robust reliable identification result local vortex detection method popular efficient easy implement produce binary output based user hard threshold however vortex global feature present challenge local detector hand global detector computationally intensive require considerable user input work propose consensus uncertainty model introduce spatial proximity enhance vortex detection result obtained using point method use four existing local vortex detector convert output fuzzy possibility value using sigmoid soft approach apply majority voting scheme enables u identify candidate vortex region higher degree confidence introduce spatial proximity based analysis discern final vortical region thus using spatial proximity coupled fuzzy input propose novel uncertainty analysis approach vortex detection use expert input better estimate system parameter result two real data set demonstrate efficacy method incorporating volumetric illumination rendering volumetric data increase visual realism lead improved spatial comprehension known spatial comprehension improved incorporating multiple light source however many volumetric illumination algorithm severe drawback dealing multiple light source drawback mainly high performance penalty memory usage tackled specialized data structure data sampling contrast paper present method enables volumetric illumination multiple light source without requiring precomputation impacting visual quality achieve goal introduce selective light update minimize required computation light setting changed discus analyze novel concept underlying selective light update demonstrate applied real data different light setting benefit using advanced illumination model volume visualization demonstrated many researcher interactive volume rendering incorporated advanced lighting achieved gpu acceleration regular volume data making volume visualization even appealing tool data exploration paper present interactive illumination strategy specially designed optimized volume visualization unstructured data basis design partial differential equation based illumination model simulate light propagation absorption scattering within volumetric medium particular two scheme introduced overcome challenge presented unstructured grid test result show added illumination effect global shadowing multiple scattering lead visually pleasing visualization also greatly enhance perception depth information complex spatial relationship feature interest volume data volume visualization enhancement introduced time unstructured grid becoming increasingly popular variety scientific simulation application animated movie popular way communicate complex phenomenon cell biology broad audience animation artist apply sophisticated illustration technique communicate story trying maintain realistic representation complex dynamic environment since hand animation time cost create goal formalize illustration technique used artist facilitate automatic creation visualization generated mesoscale particle molecular simulation technique illustrative timelapse support visual exploration complex biochemical process dynamic environment seamless temporal zooming observe phenomenon different temporal resolution visual abstraction molecular trajectory ensure observer able visually follow main actor increased visual focus event interest lens effect preserve realistic representation environment context result first user study indicate visual abstraction trajectory improves ability follow story also appreciated user lens effect increased perceived amount molecular motion environment trading traceability individual molecule paper introduce coverage opacity estimation achieve depth field effect visualizing molecular dynamic data proposed algorithm novel object approach eliminates many shortcoming state image dof algorithm based observation derived physically reference renderer coverage opacity estimation exploit semi simulate blur inherent dof effect achieves high quality dof effect augmenting atom semi shell radius proportional distance focal plane camera thus shell represents additional coverage area whose opacity varies radially based observation derived result multi dof algorithm using proposed technique becomes possible generate high quality visual result comparable achieved ground multi algorithm time obtain significant speedup essential visualizing md data enables interactive rendering paper derive underlying theory introduce coverage opacity estimation demonstrate applied real world md data order achieve dof effect analyze achieved result respect performance well quality show comparable image generated modern distributed ray tracing engine scientific phenomenon often studied collection related scalar field generated different observation phenomenon exploration data requires robust distance measure compare scalar field task identifying key event establishing correspondence feature data towards goal propose topological data structure called complete extremum graph define distance measure comparing scalar field feature manner design algorithm computing distance show application analysing time varying data contour tree topological structure associated scalar function track connectivity evolving level set function support intuitive interactive visual exploration analysis scalar function paper describes fast parallel memory efficient algorithm constructing contour tree scalar function shared memory system comparison existing implementation show significant improvement running time memory expended proposed algorithm particularly suited large datasets fit memory example contour tree scalar function defined billion vertex domain volume data efficiently constructed using le memory current architectural trend supercomputer dramatic increase number core available computational power per die power increasingly difficult programmer harness effectively high language construct simplify programming many device ease come potential loss processing power particularly cross construct recently scientific visualization package embraced language construct centering around data parallelism familiar operator map reduce gather scatter complete adoption data parallelism require central visualization algorithm revisited expressed new paradigm preserving functionality performance investment large potential payoff portable performance software base span many architecture scientific visualization application run work present method ray tracing consisting entirely data parallel primitive given extreme computational power node prevalent supercomputer believe ray tracing supplant rasterization work graphic solution scientific visualization ray tracing method relatively efficient describe performance series test also compare leading ray tracer optimized specific platform find data parallel approach lead result acceptable many scientific visualization use case key benefit providing single code base run many architecture exploring feature subspace one promising approach analyzing understanding important pattern multivariate data relying much effective enhancement manual intervention associated result depend heavily knowledge skill user performing data analysis paper present novel approach extracting feature subspace multivariate data incorporating biclustering technique approach maximally automated sense highly dimension automatically grouped form subspace effectively support exploration key idea behind approach lie new mathematical formulation asymmetric biclustering combining spherical k clustering grouping highly dimension together ordinary k clustering identifying subset data sample lower representation data feature subspace successfully visualized parallel coordinate plot project data sample correlated dimension one composite axis dimensionality reduction scheme several experimental result data analysis together discussion provided ass capability approach contextual layout preserve context data associated attribute however linear mapping cause error layout similar data point variable node may map similar region vice versa paper first unify various data layout scheme choose generalized bary coordinate plot standard way describe second propose three algorithm distance spaced lay iterative error reduction force directed adjustment reduce layout error variable variable data variable data data respectively find combination three algorithm yield large improvement layout error achieve comprehensive layout third describe interface gbc error explorer allows user explore error using variety visualization scheme combined interaction paper propose novel visual analysis method trajrank study travel behaviour vehicle along one route focus spatial distribution travel time time spent road segment travel time variation rush hour trajrank first allows user interactively select route segment several road segment trajectory passing route automatically extracted trajectory ranked road segment according travel time clustered according ranking road segment based ranking analysis provide temporal distribution view showing temporal distribution travel time ranking diagram view showing spatial variation travel time real taxi gps data present three use case informal user study show effectiveness usability method goal accurately model human class separation judgement color scatterplots towards goal propose set visual separation measure systematically combining neighborhood graph class purity function different parameterizations using machine learning framework evaluate measure based well predict human separation judgement found new measure outperform best state distance consistency measure among best measure average proportion neighbor among neighbor point target class gong dir cpt prediction accuracy better dsc also discus alternative well measure give guideline use aspect ratio plot strongly influence perception trend data arc length based aspect ratio selection demonstrated many empirical advantage previous method however still clear method work paper attempt unravel mystery exploring mathematical foundation first explain rationale method parameterization invariant follow rationale extend previous method parameterization invariant propose maximizing weighted local curvature parameterization invariant form local orientation resolution reveal theoretical connection average slope resultant vector furthermore establish mathematical connection al banking degree derive upper lower bound average absolute slope finally conduct quantitative comparison revise understanding aspect ratio selection method three aspect showing al awo rv always perform similarly m demonstrating advantage robustness rv al providing counterexample previous method produce poor result mlc work well zoomable user interface widely used static visualization many benefit however well supported animated visualization due problem change blindness information overload propose spot lens new zoomable user interface animated bubble chart tackle problem couple zooming automatic panning provides rich set auxiliary technique enhance effectiveness preliminary user study suggested besides allowing user examine detail information engaging approach exploratory analysis dynamic data paper present mathematical knot diagram interface exploit computer graphic interactive visualization multi technology enhance one intuitive experience mathematical theory knot interaction model based clever simple geometric construction named reidemeister move allows topological manipulation using rather simple move multi interface provide natural way u interact extra degree freedom characterize knot mathematical physical arithmetic property relative specialized mouse interface proposed multi interface easier intuitive learn pilot study show knot braid manipulation multi much faster efficient combine show interactive computer graphic method computer interface used construct virtual manipulatives meet challenge exploring abstract mathematical world bipartite graph model relation two different type entity applicable example describe person affiliation different social group association subject topic interest application important understand connectivity pattern among entity bipartite graph example bipartite relation person topic interest people may form group based common interest topic also grouped categorized based interested audience co method identify connectivity pattern find cluster within two type entity simultaneously paper propose interactive visualization design incorporates co method facilitate identification node cluster formed common connection bipartite graph besides highlighting automatically detected node cluster connection among visual interface also provides visual cue evaluating homogeneity bipartite connection cluster identifying potential outlier analyzing correlation node attribute cluster structure interactive visual interface allows user flexibly adjust node grouping incorporate prior knowledge domain either direct manipulation splitting merging cluster providing explicit feedback cluster quality based system learn parametrization co algorithm better align user notion node similarity demonstrate utility system present two example usage scenario real world datasets user evaluation gained increasing importance visualization research past year many case evaluation way support claim made visualization researcher unfortunately recent literature review show comparison algorithmic performance evaluation number user evaluation still low reason required amount time conduct study together difficulty involved participant recruitment result reporting could shown quality evaluation result simplified participant recruitment crowdsourcing platform make technology viable alternative lab experiment evaluating visualization time conducting reporting evaluation still high paper propose software system integrates conduction analysis reporting crowdsourced user evaluation directly scientific visualization development process proposed system researcher conduct analyze quantitative evaluation large scale evaluation user interface mouse click thus becomes possible perform iterative evaluation algorithm design potentially lead better result compared time consuming user evaluation traditionally conducted end design process furthermore system built around centralized database support easy reuse old evaluation design reproduction old evaluation new additional stimulus driving challenge scientific visualization research describe system design consideration made design process demonstrate system conducting three user evaluation published visualization literature cosmological simulation produce multitude data type whose large scale make difficult thoroughly explore interactive setting one aspect particular interest scientist evolution group dark matter particle described merger tree however order fully understand subtlety merger tree data type derived simulation must incorporated well work develop novel interactive linked visualization system focus simultaneously exploring dark matter halo hierarchical evolution corresponding particle data quantitative information employ parallel remote renderer local merger tree selection tool user analyze large data set interactively allows scientist ass simulation code understand inconsistency extracted data intuitively understand simulation behavior scale demonstrate effectiveness system set case study large cosmological data hacc accelerated cosmology code simulation framework paper present novel visual analysis tool ensemblegraph aim helping scientist understand spatiotemporal similarity across run time ensemble simulation data abstract input data graph node represents region similar behavior across run node adjacent time frame linked region overlap spatially visualization graph combined multiple view showing detail enables user explore select compare extracted region similar behavior driving application paper study regional emission influence tropospheric ozone based ensemble simulation conducted different anthropogenic emission absence using mozart demonstrate effectiveness method visualizing mozart ensemble simulation data evaluating relative regional emission influence tropospheric ozone concentration nuclear engineering understanding safety margin nuclear reactor via simulation arguably paramount importance predicting preventing nuclear accident therefore crucial perform sensitivity analysis understand change model input affect output modern nuclear simulation tool rely numerical representation sensitivity information inherently lacking visual encoding offering limited effectiveness communicating exploring generated data paper design framework sensitivity analysis visualization multidimensional nuclear simulation data using partition topology regression model report efficacy rely established morse regression technique allows u partition domain monotonic region easily interpretable linear model used ass influence input output variability underlying computation augmented intuitive interactive visual design effectively communicate sensitivity information nuclear scientist framework deployed multipurpose probabilistic risk assessment uncertainty quantification framework raven analysis virtual control environment evaluate framework using simulation dataset studying nuclear fuel performance vector field topology one powerful flow visualization tool break huge amount data compact sparse easy read description little information loss suffers one main drawback though definition critical point foundation vector field topology highly dependent frame reference paper propose consider every point critical point locally adjust frame reference persistent one mean extremum determinant jacobian result extraction one well frame reference simultaneous visualization dominating frame reference different area flow field could individually perceived observer traveling along critical point show important one present novel high access dependency model efficient pathline computation unsteady flow visualization taking longer access sequence account model sophisticated data access pattern particle tracing method greatly improves accuracy reliability data access prediction work high access dependency calculated tracing uniformly seeded pathlines forward backward direction preprocessing stage effectiveness approach demonstrated parallel particle tracing framework high data prefetching result show method achieves higher data locality hence improves efficiency pathline computation comparing time surface different integration time point different seeding area provide valuable insight transport phenomenon fluid flow comparative study challenging due often convoluted shape surface propose new approach comparative flow visualization based time surface exploit idea embedding surface carefully designed reformed visualization space embedding enables new opportunity comparative flow visualization present three different strategy comparative flow visualization take advantage reformation reforming time surface mitigate occlusion issue devote also third dimension visualization space comparative aspect visualization approach effective variety flow study case direct comparison individual time surface reveals small scale difference fine detail fluid motion concurrent study multiple surface family enables identification comparison prominent motion pattern work developed close collaboration expert fluid dynamic assessed potential usefulness approach field propose longest common subsequence approach compute distance among vector field ensemble measuring many common block ensemble pathlines pas lcss distance defines similarity among vector field ensemble counting number shared domain data block compared traditional method pointwise euclidean distance dynamic time warping distance proposed approach robust outlier missing data sampling rate pathline timesteps taking advantage smaller reusable intermediate output visualization based proposed lcss approach reveals temporal trend data low storage cost avoids tracing pathlines repeatedly evaluate method synthetic data simulation data demonstrating robustness proposed approach study cavity channel molecular structure crucial step understanding function biomolecules current tool technique extracting structural feature sensitive uncertainty atomic position radius paper study problem cavity extraction biomolecules taking account uncertainty propose approach connects user cavity computing optimal conduit within region occupied molecule conduit computed using topological representation occupied empty region guaranteed satisfy well defined geometric optimality criterion visualization set cavity multiple linked view serf useful interface interactive extraction stable cavity demonstrate utility proposed method successfully identifying biologically significant pathway molecular cavity using several case study reactivity biomolecular structure highly influenced structural feature thus studying feature along exploration dynamic behavior help understand process ongoing living cell reached visual representation process visualization one natural way convey information however none currently available technique provides biochemist intuitive real representation dynamic movement molecule precise geometrical based extraction structural feature performed instantly paper introduce technique enabling user compute also visualize molecular surface along inner void obtain better insight molecule technique enables visualize molecular surface transparently opacity adjusted changing user parameter order enhance perception surface inner void integrated algorithm run real give user big variety exploration possibility importance approach even amplified respect fact currently size molecular dynamic simulation increasing dramatically offline rendering thus becomes impracticable usability technique evaluated domain expert exploration atomic trajectory fundamental understanding characterizing complex chemical system important many application instance new insight mechanism ionic migration catalytic material could lead substantial increase battery performance new statistical measure called relative distribution proposed understand complex motion whether brownian ballistic diffusive relative distribution represented collection histogram currently created slow offline process making parameter exploration tedious time task furthermore resulting plot hide uncertainty data visualization result rastered printed fixed resolution histogram misleading present new analysis tool exploration atomic trajectory combine interactive histogram visualization uncertainty information data plotting error also linked interactive display trajectory tool enables holistic exploration trajectory previously feasible potential significant scientific impact collaboration domain expert deployed tool ta analyze molecular dynamic simulation lithium diffusion user found tool significantly accelerates exploration process used validate number previously unconfirmed hypothesis subspace analysis increasingly become preferred method clustering high data visually interactive exploration subspace cluster cyclic process every meaningful discovery motivate user subspace provide improved clustering result reveal relationship among cluster hardly coexist original subspace however combination dimension original subspace always effective finding expected subspace study present approach enables user reconstruct new dimension data projection subspace preserve interesting cluster information reconstructed dimension included analytical workflow original dimension help user construct target subspace clearly display informative cluster structure also provide visualization tool assist user exploration subspace cluster utilizing dimension reconstruction several case study synthetic real data set performed prove effectiveness approach lastly evaluation approach conducted via expert review domain epidemiology logistic regression modeling widely used explain relationship among explanatory variable dichotomous outcome variable however logistic regression modeling face challenge overfitting confounding multicollinearity large number explanatory variable example birth defect study presented paper variable selection building high quality model identify risk factor hundred pollutant variable difficult address problem propose novel visual analytics approach logistic regression modeling high datasets leverage traditional modeling pipeline providing intuitive visualization inspecting statistical indicator relationship among variable seamless effective dimension reduction pipeline selecting variable inclusion high quality logistic regression model fully working prototype approach developed successfully applied birth defect study illustrates effectiveness efficiency application insurance policy study feedback domain expert demonstrate usefulness word cloud popular method visualize frequency word textual data nowadays many text data set flickr tag geo important spatial component however existing automated method generate word cloud unable incorporate spatial information introduce geo word cloud word cloud capture frequency also spatial relevance word input set location one geographic region several text label per location aggregate word frequency according point cluster employ greedy strategy place appropriately sized label without overlap close possible corresponding location spatial shape geographic region corresponding label experimentally explore trade concerning location label relative size number spatial cluster resulting word cloud visually pleasing low error term relative scaling locational accuracy word using small number cluster per label present new approach analysis large image collection argue categorization play important role process label image end result also exploration furthermore increase effectiveness efficiency categorization process enable use available metadata treated multivariate data identified image attribute category important aspect integrated system called iclic system consists four view connected selection image central action approach using minimalist interface standard metaphor user enabled use system short time system enables complex query natural way deal collection containing image metadata attribute confirmed two evaluation cycle domain expert stored procedure used database system process aggregate data hundred stored procedure often form complex process network documented hidden dependency difficult understand maintain debug paper introduces novel approach support task visually comparing specific process run run process visualization based force node diagram arranged timeline color coding histogram trend chart used highlight temporal deviation approach implemented interactive web application used professional database developer solving realistic maintenance debugging task feedback expert user confirms usefulness practical relevance approach movement data collected gps technology increasingly common difficult visualize due overplotting occlusion movement displayed map additional challenge extraction useful higher information meeting derived raw movement data present design study movementslicer tool visualizing place visited behavior individual actor also meeting multiple actor first present taxonomy visualization movement data consider task support analyzing movement data especially meeting multiple actor argue gantt chart many advantage understanding movement meeting small group moving entity present design gantt chart nest people within location location within people along vertical axis show time along horizontal axis row gantt chart sorted activity level filtered using weighted adjacency matrix showing meeting people empty time interval gantt chart automatically folded smoothly animated transition yielding multi view case study demonstrate utility prototype geo microblog data cover billion movement pattern global local scale understanding pattern could guide urban traffic planning help coping disaster situation present visual analytics system investigate travel trajectory people reconstructed microblog message analyze seasonal change event validate movement pattern data source contribute highly interactive visual comparison method normalize contrast trajectory well density map within single view also compute adaptive hierarchical graph trajectory abstract individual movement higher structure specific challenge tackle among others spatio sparsity data volume data varying region diverse mix mean transportation applicability approach presented three case study many real world information represented graph set node interconnected multiple type relation called edge layer social network biological data edge bundling technique proposed solve cluttering issue standard graph effort done deal similar issue multilayer graph multilayer graph scenario clutter induced large amount edge problem also fact different type edge overlap making useless final visualization paper introduce new multilayer graph edge bundling technique firstly produce preliminary edge bundling independently different edge layer deal specificity multilayer graph one type edge routed bundle proposed visualization tested real world case study outcome point ability proposal discover pattern present data origin matrix useful tool traffic transport analysis kind matrix increasing volume complexity due availability new technology gather new information project present novel way representing origin matrix using connection barchart two coordinated map main goal use connection barchart center interactive visualization providing overview dataset interaction information specific origin node origin matrix shown connection barchart coordinated map giving user information origin destination direction volume flow additionally evaluation user conducted test usefulness usability interactive visualization using real data rbt system bogota colombia rising trend moocs attracted wide ranging research interest among existing study related moocs focus individual study behavior evaluation analysis click stream video behavior exploration etc course design purpose however addition traditional course material moocs also provide interactive user forum encourage student seek help peer endows course social network formation interaction thus present networkseer help evaluate mooc student use forum networkseer visualizes interaction forum including interaction happen also enables filtering un group case study conducted demonstrate usefulness curve long used graph visualization increased popularity recent year curve mainly used two purpose one increase readability enhance visual aesthetic pleasingness although curve visually pleasing introduction curve graph drawing increase readability automatically attempt made investigate usability curved drawing however result effect curve per se human graph comprehension conclusive paper present user study examine effect curve introduced remove crossing twenty participant recruited perform typical graph reading task task performance user preference data collected analysis result indicate curve useful alternative crossing present straight drawing finding study also discussed along future research activity paper visualizing set isosurfaces scalar ensemble field difficult task due inherent occlusion effect yet often required analyze uncertainty represented ensemble paper present novel visualization technique ensemble isosurfaces based screen silhouette using silhouette displayed information reduced avoid occlusion yet major shape surface maintained approach preserve spatial coherence make assumption underlying surface distribution providing additional mechanism picking clustering cutting animation enable user explore ensemble surface interactively visualizing similarity difference among ensemble isosurfaces challenging problem mainly isosurfaces cannot displayed together time ensemble isosurfaces visualizing spatial difference among surface essential get useful insight individual ensemble simulation affect different isosurfaces propose scheme visualize spatial variation isosurfaces respect statistically significant isosurfaces within ensemble understanding variation among ensemble isosurfaces different spatial region helpful analyzing influence different ensemble run spatial domain regard propose isosurface based clustering scheme divide spatial domain region high low isosurface variation demonstrate efficacy method successfully applying real ensemble data set ocean simulation experiment weather forecast streamline technique play important role visualizing analyzing uncertain steady vector field challenging problem generate accurate streamlines uncertain vector field due global uncertainty transportation work present novel probabilistic method streamline computation uncertain steady vector field using bayesian framework framework streamline modeled state space model capture spatial coherence integration step uncertainty local distribution using conditional prior density likelihood function approximate posterior distribution possible trace originating given seed position set weighted sample iteratively updated streamlines higher likelihood derived qualitatively quantitatively compare method alternative method different type flow field data set method generate possible streamlines higher certainty hence accurate flow trace assessment using motion visualization polarized conjecture motion may effective comparative visualization applied properly especially dealing large amount multi data designed visualization represent driver behavior series user study conducted verify adding motion static visualization help user make comparison separate drastically different behavior result show adding motion indeed lead shorter completion time le cognitive workload percentage individual frequently using smartphones work life increasing steadily interaction individual smartphones produce large amount usage data contain rich information smartphone owner usage habit daily life paper visual analytic tool proposed discover understand individual behavior pattern smartphone usage data four cooperated visualization view many interaction provided tool visually explore temporal feature various interactive event smartphones user hierarchical association among event type detailed distribution massive event sequence case study plenty interesting pattern discovered analyzing data two smartphone user different usage style advancement motion sensing input device enabled collection multivariate time body movement data analyzing type data challenging due large amount data task mining interesting temporal movement pattern address problem propose interface visualize analyze body movement data visualization enables user navigate explore evolution movement time different movement area also propose clustering method based hierarchical clustering group similar movement pattern proposed visualization illustrated case study demonstrates ability interface analyze body movement stacked graph widely used represent multiple time series simultaneously show change individual value aggregation time however number time series becomes large layer representing time series small value take small proportion stacked graph making hard trace result challenging analyst detect correlation individual layer aggregation find trend similarity difference layer solely stacked graph paper study correlation individual layer aggregation time series data presented stacked graph focusing local region within given time interval specifically present stac interactive visual analytics system help analyst gain insight correlation stacked graph preserving original stacked shape link stacked graph auxiliary view facilitate analysis correlation time series data case study based real dataset demonstrates effectiveness system gaining insight time series data analysis facilitating various analytical task word cloud widely used present content theme text summary visualization paper propose new semantic word cloud taking account word semantic meaning distributed word representation applied accurately describe semantic meaning word word similarity graph constructed based semantic distance word lay word compact aesthetic manner word interaction introduced guide user fast read understand text apply proposed word cloud user generated review different field demonstrate effectiveness method paper present comparative study personal visualization bibliographic data consider three design egocentric visualization node diagram adjacency matrix botanical tree depict one academic career term publication record case study conducted compare effectiveness resulting visualization conveying particular aspect researcher bibliographic record based study find node diagram better revealing overall distribution certain attribute adjacency matrix convey information le clutter botanical tree visually attractive provide best glance characterization mapped data mapping data tree feature must carefully done derive expressive visualization online bookstore highly thrived changed consumer behavior year however customer go online bookstore specific target one reason current web interface usually complex cluttered user browse addition current visualization interface display result associated single attribute thus requiring user interact intensively find target inspired user experience physical bookstore present bookwall online bookstore interface comprises two component category map wall view enabling user find target efficiently releasing user burden complicated operation specifically category map produce map map look providing overview cluster neighborhood book category wall view enables displaying query result satisfying dual query attribute simultaneously result show bookwall provide user favourable alternative visualization genome wide association study examine genetic variant different individual detect variant associated specific disease genome project collaborative research effort sequence genome least participant different ethnicity establish detailed summary human genetic variation kinship information measure individual ancestor relationship within considered population study design kinship data visualization allowing expert discover anomaly gwas data visual analysis genome project kinship data reveals inconsistency call deeper analysis data quality within project share challenge lesson learned designing exhibit prototype encourages museum visitor learn marine animal behavior interactive visualization data exploration intent visitor draw comparison animal behavior similarly scientist would make insight discovery effort designed set visual encoding around tagging pelagic predator data set create appropriate abstraction rich complex field data incorporated multiple external representation tangible user interface provide complementary representation data promote self formative evaluation identify strength weakness prototype design evaluation result suggest progressing right direction observed public making comparison inference still require design iteration improve visualization exhibit