code and link to data for mapping spatial patterns in conservation research

mappingConservationResearch.R

# mapping conservation research

#load libraries
library(geosphere)
library(dplyr)
library(StandardizeText)
library(ggmap)
library(maps)
library(rvest)
library(rworldmap)
library(RColorBrewer)
library(geoR)

#read raw data
abstractsRaw <- read.csv("https://raw.githubusercontent.com/luisDVA/codeluis/master/abstracts.csv",stringsAsFactors = FALSE) 
# manual fix for some country names, and standardize to Penn World Tables names
abstracts <- abstractsRaw %>% 
  filter(mappable=="Yes") %>% select(authorCountry,studyCountry) %>% 
  mutate_each(funs(gsub("USA","United States",.))) %>% 
  mutate_each(funs(gsub("UK","United Kingdom",.))) %>% 
  mutate_each(funs(standardize.countrynames(standard = "pwt", suggest="prompt",.)))

# create table for plotting connections
connections <- abstracts %>% filter(authorCountry!=studyCountry)

# expanding multiple connections
absMult <- abstractsRaw %>% 
  filter(multipleConns=="Yes") %>% 
  select(authorCountry,studyCountry)
# make new DF with additional connections from multi-study area papers
y<-strsplit(as.character(absMult$studyCountry) , ", ", fixed=TRUE)
moreConnections <- data.frame(authorCountry = rep(absMult$authorCountry, sapply(y, length)),studyCountry= unlist(y)) %>% 
  mutate_each(funs(gsub("USA","United States",.))) %>% 
  mutate_each(funs(gsub("UK","United Kingdom",.))) %>% 
  mutate_each(funs(standardize.countrynames(standard = "pwt", suggest="prompt",.)))

# merge both DFs
connections <- bind_rows(connections,moreConnections)

# tally in-country research
localRes <- abstracts %>% filter(authorCountry==studyCountry) %>% 
  count(authorCountry)

#Scrape capital cities table from web
countriesTab <- html("http://geographyfieldwork.com/WorldCapitalCities.htm")

#get names and capitals
cap_table <-  countriesTab %>% 
  html_node(".sortable") %>%
  html_table(., fill = T) %>% 
  rename(Capital=`Capital City`)


# capital cities to geocode

capAC <- merge(connections,cap_table,by.x="authorCountry",by.y="Country",all.x=T)
capSC <- merge(capAC,cap_table,by.x="studyCountry",by.y="Country",all.x=T)
capitalsC <- select(capSC,authLoc=Capital.x,studLoc=Capital.y)

# geocode and jitter author locations
coordsAuthLoc <- mutate_geocode(capitalsC,authLoc,output="latlon",source="google")%>% 
                 select(authLoc,studLoc,latAut=lat,lonAut=lon)
coordsAuthLoc[,3:4] <- jitterDupCoords(coordsAuthLoc[,3:4],max=0.7)

# geocode and jitter study locations
coordsAuthStudLoc <- mutate_geocode(coordsAuthLoc,studLoc,output="latlon",source="google") %>% 
                     select(authLoc,studLoc,latAut,lonAut,latStud=lat,lonStud=lon)
coordsAuthStudLoc[,5:6] <- jitterDupCoords(coordsAuthStudLoc[,5:6],max=0.7)

#join "local research" table to a coarse resolution map
localresearch <- joinCountryData2Map(localRes, joinCode="NAME", nameJoinColumn="authorCountry")


#create a map-shaped window
mapDevice('x11')

#plot
par(bg="grey15")
mapParameters <- mapCountryData(localresearch, nameColumnToPlot="n", catMethod="fixedWidth",
               borderCol="grey11", oceanCol="grey15",missingCountryCol = "grey8",addLegend = F,
               mapTitle = "Conservation research",
               colourPalette = c("#3182BD", "#00004d"))
do.call(addMapLegend,c(mapParameters,legendWidth = 0.5))


# Great circle lines to connect points

for (i in 1:length(coordsAuthStudLoc$lonAut)) {
  inter <- gcIntermediate(c(coordsAuthStudLoc$lonAut[i], coordsAuthStudLoc$latAut[i]), c(coordsAuthStudLoc$lonStud[i], coordsAuthStudLoc$latStud[i]), n=500, addStartEnd=TRUE, breakAtDateLine=TRUE)
  if (length(inter) > 2) {
    lines(inter,col="white",lwd=0.5)
  } else {
    lines(inter[[1]],col="white",lwd=0.5)
    lines(inter[[2]],col="white",lwd=0.5)
  }
}                


#overlay points
points(coordsAuthStudLoc$lonAut,coordsAuthStudLoc$latAut, pch=16, cex=0.8, col="orange")
points(coordsAuthStudLoc$lonStud,coordsAuthStudLoc$latStud, pch=16, cex=0.8, col="blue")