How to convert state name to latitude and longitude to create choropleth with ggvis in R?

Question

I want to create a choropleth for income values in each state with package ggvis. Below is my data. I think latitude and longitude columns are required for mapping the data. Anyone knows that how can I convert the state name to latitude and longitude? Thanks a lot!

mapdata<-data.frame(
 state=c("alabama","alaska","arizona","arkansas","california","colorado","connecticut","delaware","florida","georgia","hawaii","idaho","illinois","indiana","iowa","kansas","kentucky","louisiana","maine","maryland","massachusetts","michigan", "minnesota","mississippi","missouri","montana","nebraska","nevada","new hampshire","new jersey","new mexico","new york","north carolina","north dakota","ohio","oklahoma", "oregon","pennsylvania","rhode island","south carolina","south dakota","tennessee","texas","utah","vermont","virginia","washington","west virginia","wisconsin","wyoming"),
  income=runif(50,min=100,max=9000))

Answer 1

Here's one way. I'm using ggplot rather than ggVis but this should get you started.

library(raster)       # for getData(...)
library(ggplot2)
library(data.table)
usa   <- getData('GADM',country='USA',level=1)         # shapefile of US states
shp   <- usa[(!usa$NAME_1 %in% c("Alaska","Hawaii")),] # remove AK, HI for this example
gg.dt <- setDT(fortify(shp))               # convert shapefile to format ggplot can use
# merge with attribute table
gg.dt <- gg.dt[setDT(cbind(id=rownames(shp@data),shp@data)),on="id"]  
gg.dt[,state:=tolower(NAME_1)]             # convert state names to lower case                                     
gg.dt <- gg.dt[setDT(mapdata),on="state"]  # merge with mapdata

ggplot(gg.dt, aes(x=long, y=lat, group=group,fill=revenue)) + 
  geom_polygon(color="grey50", size=0.1)+
  scale_fill_gradientn(colours=rev(heat.colors(10)))+
  coord_map()

So here is a ggvis solution (probably should be a separate answer, but WTH). It turns out that the shapefile above is very high resolution and ggvis just can't deal with that large a file. So here we download a low-res shapefile of US state boundaries, merge using data.frames (which is adequately fast with the low-res shapefile), and then render using ggvis.

library(rgdal)   # for readOGR(...)
library(ggplot2) # for fortify(...)
library(ggvis)
# load low resolution US state shapefile (1:20MM)
url   <- "http://www2.census.gov/geo/tiger/GENZ2014/shp/cb_2014_us_state_20m.zip"
tf    <- tempfile()
td    <- tempdir()
download.file(url,tf, mode="wb")   # download shapefile archive of US state boundaries
unzip(tf, exdir=td)                # unzip into directory td

usa <- readOGR(dsn=td, layer="cb_2014_us_state_20m")
shp <- usa[(!usa$STUSPS %in% c("AK","HI")),] # remove AK, HI for this example

gg.df       <- fortify(shp)                     # convert shapefile to format ggvis can use
gg.df       <- merge(gg.df,cbind(id=rownames(shp@data),shp@data),by="id")   # merge with attribute table
gg.df$state <- tolower(gg.df$NAME)              # convert state names to lower case                                     
gg.df       <- merge(gg.df,mapdata,by="state")  # merge with mapdata
gg.df       <- gg.df[order(gg.df$order),]       # reset to original order

gg.df %>%
  group_by(group) %>%
  ggvis(~long, ~lat)  %>%
  layer_paths(fill= ~revenue)

There's a decent blog post on creating choropleth maps using ggvis here. If you want custom colors (as in the earlier answer), see the post.