SpatialAnalysis.Rmd

---
title: "Spatial Analysis"
output: html_notebook
---

```{r}

library(tidyverse)
library(tidyverse)
library(gghighlight)
library(purrr)
library(plotly)
library(cowplot)
library(scales)

```

```{r}

readSpatialFile<-function (fname,fpath){        # read and process a single file
#  Function to read in Work/PlayInfections.dat' type files (many columns, one for each class only of infecteds, Time is row number)
  fname=str_c(fpath,fname)
  read.table(file=fname,sep =' ') %>% # make long data frame
  mutate(Time=row_number()) %>% # count up time, as row number
  pivot_longer(-Time) %>% 
  mutate(Ward=as.integer(str_remove(name,'V'))) %>% 
  select(-name)->inf_long # rename name to Ward integers for easier matching
  return(inf_long)
}


#readSpatialFile(spatialfiles[1],pathname)->tmp # How  to use the above two functions. 

dataS <- tibble(filename = spatialfiles) %>% # create a data frame
                                                  # holding the file names
  mutate(file_contents = map(filename,            # read files into a new data column
                             ~readSpatialFile(., pathname))) %>% # use the function written above to process files
  mutate(Location=(str_split(filename,'/',simplify = TRUE)[,1])) %>% 
  mutate(Run=as.numeric(str_split(filename,'/',simplify = TRUE)[,2])) %>% 
  select(-filename) %>% 
  unnest(cols=file_contents)


```

```{r functions}
lookupfile='~/GitHub/MetaWards/data/2011/WardsProcessing/WardLookupRegionCity.csv'

readlookupfile<-function (fname){
df <-   read_csv(lookupfile,col_name=TRUE,cols(
  WD11CD = col_character(),
  WD11NM = col_character(),
  WD11NMW = col_character(),
  CMWD11CD = col_character(),
  CMWD11NM = col_character(),
  CMWD11NMW = col_character(),
  IND = col_double(),
  LAD11CD = col_character(),
  LAD11NM = col_character(),
  LAD11NMW = col_character(),
  FID = col_double(),
  Region = col_character(),
  City = col_character()
)
)
return(df)
}

wardlookupregionCity <-readlookupfile(lookupfile)

wardlookupregionCity %>% 
  inner_join(.,inf_long, by=c('FID'='Ward'),all.y=T,all.x=F) %>%
  group_by(City,Time)%>% 
  summarise(Cases=sum(value)) -> city_inf

TimeToPeakCities<-function (df,cities){
  df %>% 
    filter(City%in%cities) %>% 
    group_by(City) %>% 
    summarise(TimeToPeak=which.max(Cases))->d
  return(d)
}


ExtractTimeToPeakCities <- function(fname,path,cities,lookupfile){
  wardlookupregionCity <- readlookupfile(lookupfile)
  readSpatialFile(fname,path)->inf_long
  
  wardlookupregionCity %>% 
    inner_join(.,inf_long, by=c('FID'='Ward'),all.y=T,all.x=F) %>%
    group_by(City,Time)%>% 
    summarise(Cases=sum(value)) -> city_inf
  
  TimeToPeakCities(city_inf,cities)->d
  return(d)
}


cities=c('London','Liverpool','Birmingham','Sheffield','Bristol','Cardiff','Manchester')

ExtractTimeToPeakCities(spatialfiles[1],pathname,cities,lookupfile)



```

now  do it for all of them

```{r}

pathname='~/GitHub/MetaWards/Runs/29_02/'
dir(pathname,pattern='ForMattData.dat',recursive=TRUE)->spatialfiles


# FromToPeak <- tibble(filename = spatialfiles) %>% # create a data frame
#                                                   # holding the file names
#   mutate(timetopeaks = map(filename,            # read files into a new data column
#                              ~ExtractTimeToPeakCities(.,pathname,cities,lookupfile))) %>% # use the function written above to process files
#   mutate(Location=(str_split(filename,'/',simplify = TRUE)[,1])) %>% 
#   mutate(Run=as.numeric(str_split(filename,'/',simplify = TRUE)[,2])) %>% 
#   select(-filename) %>% 
#   unnest(cols=timetopeaks)

# write_csv(path='~/GitHub/MetaWards/Runs/29_02/FromToPeak.csv',FromToPeak)

FromToPeak=read_csv(file='~/GitHub/MetaWards/Runs/29_02/FromToPeak.csv')

```


```{r}

FromToPeak %>% 
  filter(City!='Newcastle upon Tyne') %>% 
  mutate(R0=ifelse(Run<10,'1.95','2.4'))%>% 
  mutate(WeeksToPeak=floor(TimeToPeak/7)) %>% 
  ggplot(aes(y=WeeksToPeak,x=City))+
  geom_boxplot()+
  facet_grid(R0~Location,scales = 'free_y')+
  theme(axis.text.x = element_text(angle = 90))
```
```{r}
library(viridis)

mid=mean(FromToPeak$TimeToPeak/7)

FromToPeak %>% 
  filter(City!='Newcastle upon Tyne') %>% 
  mutate(R0=ifelse(Run<10,'R0=1.95','R0=2.4'))%>% 
  mutate(Location=as.factor(Location)) %>% 
  mutate(City=as.factor(City)) %>% 
  group_by(Location,City,R0) %>% 
  summarise(MeanTimeToPeak=mean(TimeToPeak/7))%>% 
  ggplot(aes(x=Location,y=City,colour=MeanTimeToPeak))+
  geom_point(size=7)+
  facet_wrap(~R0)+
  scale_color_gradient2(midpoint=mid, low="darkblue", mid="red",
                     high="white", space ="Lab" )+
  theme(axis.text.x = element_text(angle = 90))+
  ylab('Destination city')+
  xlab('Seeding city')





```


```{r}

# path='~/GitHub/MetaWards/Runs/29_02/Generalised/'
# dir(path,pattern='ForMattData.dat',recursive=TRUE)->generalisedSF
# 
# FromToPeakGeneralised<-
#   tibble(filename = generalisedSF) %>% # create a data frame
#                                                    # holding the file names
#    mutate(timetopeaks = map(filename,            # read files into a new data column
#                               ~ExtractTimeToPeakCities(.,path,cities,lookupfile))) %>% # use the function written above to process files
# #   mutate(Location='Generalised') %>% 
#    mutate(Run=as.numeric(str_split(filename,'/',simplify = TRUE)[,1])) %>% 
#    select(-filename) %>% 
#    unnest(cols=timetopeaks)

#write_csv(path='~/GitHub/MetaWards/Runs/29_02/FromToPeakGeneralised.csv',FromToPeakGeneralised)

read_csv(path='~/GitHub/MetaWards/Runs/29_02/FromToPeakGeneralised.csv')

FromToPeakGeneralised %>% 
  filter(City!='Newcastle upon Tyne') %>% 
  filter(Run!='20') %>% 
  mutate(R0=ifelse(Run<10,'R0=1.95','R0=2.4'))%>% 
  mutate(WeeksToPeak=floor(TimeToPeak/7))%>% 
  ggplot(aes(y=WeeksToPeak,x=City))+
  geom_boxplot()+
  facet_grid(~R0)+
  theme(axis.text.x = element_text(angle = 90))+
  xlab('Destination city')+
  ylab('Number of weeks to peak')
  
ggsave(filename = '~/GitHub/MetaWards/ForSpiM/02_03/Figure2a.png')
```


```{r}


path='~/GitHub/MetaWards/Runs/29_02/Generalised/'
dir(path,pattern='ForMattData.dat',recursive=TRUE)->generalisedSF
 
readSpatialFile(generalisedSF[15],path)->inf_long
  
wardlookupregionCity %>% 
    inner_join(.,inf_long, by=c('FID'='Ward'),all.y=T,all.x=F) %>%
    group_by(City,Time)%>% 
    summarise(Cases=sum(value)) -> city_inf


GeneralisedAll<-
    tibble(filename = generalisedSF) %>% # create a data frame
                                                    # holding the file names
    mutate(spatialepidemics = map(filename,            # read files into a new data column
                               ~readSpatialFile(.,path))) %>% # use the function written above to process files
# #   mutate(Location='Generalised') %>% 
    mutate(Run=as.numeric(str_split(filename,'/',simplify = TRUE)[,1])) %>% 
    select(-filename) %>% 
    unnest(cols=spatialepidemics)



wardlookupregionCity %>% 
    inner_join(.,GeneralisedAll, by=c('FID'='Ward'),all.y=T,all.x=F)%>%
    group_by(City,Run,Time)%>% 
    summarise(Cases=sum(value)) -> city_inf_all
#write_csv(path='~/GitHub/MetaWards/Runs/29_02/City_Inf_All_Generalised.csv',city_inf_all)

```

```{r}

%>% 
  filter(Time>20) read_csv(file='~/GitHub/MetaWards/Runs/29_02/City_Inf_All_Generalised.csv')->city_inf_all

cities2=c('London','Cardiff','Sheffield')

city_inf_all %>% 
  filter(Time<200) %>% 
  filter(City%in%cities2) %>% 
  filter(Run%in%c(0,10)) %>% 
  group_by(City,Run) %>% 
  mutate(PropInfectious=Cases/sum(Cases),TimeToPeak=which.max(Cases))%>%
  mutate(TimeFromPeak=Time-TimeToPeak) %>% 
  mutate(R0=ifelse(Run<10,'R0=1.95','R0=2.4')) %>% 
  ggplot(aes(x=Time,y=PropInfectious))+
  geom_line(aes(colour=City),alpha=0.9,size=1)+
  facet_wrap(~R0)+
  theme_minimal()
```


```{r}
readSpatialFile(generalisedSF[15],path)->inf_long
  
wardlookupregionCity %>% 
    inner_join(.,inf_long, by=c('FID'='Ward'),all.y=T,all.x=F) %>%
    group_by(City,Time)%>% 
    summarise(Cases=sum(value)) -> city_inf

city_inf %>% 
  group_by(City) %>% 
  filter(City%in%cities2) %>% 
  filter(Time<250, Time>80) %>% 
  mutate(PropInfectious=Cases/sum(Cases),TimeToPeak=which.max(Cases))%>%
  mutate(TimeFromPeak=Time-TimeToPeak) %>% 
  ggplot(aes(x=Time,y=PropInfectious))+
  geom_line(aes(colour=City),alpha=0.9,size=1)+
 # facet_wrap(~R0)+
  theme_minimal()



city_inf %>% 
  filter(City%in%cities) %>% 
  group_by(City) %>% 
  summarise(Peak=which.max(Cases))
  

```


## By Week

```{r}
# by week
city_inf_all %>% 
  filter(City%in%cities2) %>% 
  filter(Run%in%c(0,10)) %>% 
  group_by(City,Run) %>% 
  mutate(PropInfectious=Cases/sum(Cases),TimeToPeak=which.max(Cases))%>%
  mutate(TimeFromPeak=Time-TimeToPeak) %>% 
  ungroup() %>% 
  group_by(City,Run,WeekFromPeak=floor(TimeFromPeak/7)) %>% 
  summarise(WeeklyInf=sum(PropInfectious))%>% 
  mutate(R0=ifelse(Run<10,'R0=1.95','R0=2.4')) %>% 
  ggplot(aes(x=WeekFromPeak,y=WeeklyInf))+
  geom_line(aes(colour=City))+
  facet_wrap(~R0)



```