The repository contains geospatial analyses scripts to show the local relationships and dependencies of particulate matter measurement data with semivariances. Also, temporal relationships as season or weekly-dependent differences are analyzed here.
To explore kriging and the analyses, first run a jupyter notebook.
jupyter notebook
Then the kriging process can be explored
get_raw_data >> distance_matrix >> variogram_cloud >> empirical_variogram >> semivariogram
[grid, get_raw_data, distance_matrix, semivariogram] >> kriging >> result
in the analysis.ipynb with the parameters start-time [UTC] (time for analysis), bounding-box [geojson-like] (spatial bounding box), distance-bins [m] (size of distance bins), max-range [UTC] (maximum distance for PM correlation in kriging) and target-grid [m] (calculation grid size).
When different parameters (e.g. time ranges) are analysed, also the differences between the dependencies (semivariances) can be explored with statistical significance tests in statistical_test.ipynb.
Alternatively, you can run and parametrize both notebooks with papermill.
analysis.ipynb:
papermill analysis.ipynb analysis_<start-time>_<bounding-box>_<distance-bins>_<max-range>.ipynb -p START_TIME <start-time> -p BOUNDING_BOX <bounding-box> -p DISTANCE_BINS <distance-bins> -p MAX_RANGE <max-range> -p TARGET_GRID <target-grid>
statistical_test.ipynb:
papermill statistical_test.ipynb statistical_test_<start-time1>_<bounding-box1>_<distance-bins1>_<max-range1>_<start-time2>_<bounding-box2>_<distance-bins2>_<max-range2>.ipynb -p START_TIME1 <start-time1> -p BOUNDING_BOX1 <bounding-box1> -p DISTANCE_BINS1 <distance-bins1> -p MAX_RANGE1 <max-range1> -p START_TIME2 <start-time2> -p BOUNDING_BOX2 <bounding-box2> -p DISTANCE_BINS2 <distance-bins2> -p MAX_RANGE2 <max-range2>