Data visualization: Maps using ggplot2 and facet layout

Question

I would like to create a conditional micromap using R. Basically I would like to create a facet (grid layout) for four different variables in the example below p_1,p_2,q_1,q_2) and plot each states map with color coded 1 for blue and 0 for green.

Below is the sample code. The data that is to be used for color coding is "mydata", 0 for green and 1 for blow for each of the variable p_1,p_2,q_1,q_2 How would I accomplish this in using ggplot.

library(ggplot2)
library(maps)
library(scales) # for function alpha()
us.dat <- map_data("state")

ggplot(us.dat, aes(x=long, y=lat, group=group)) + geom_polygon(fill="grey65", colour = alpha("white", 1/2), size = 0.2) + 
  theme_bw() + theme(legend.position = "none", text = element_blank(), line = element_blank()) + coord_map("polyconic") 

# create random data

states <- unique(us.dat$region)
p_1 <- sample(0:1,49,replace=T)
p_2 <- sample(0:1,49,replace = T)
q_1 <- sample(0:1,49,replace=T)
q_2 <- sample(0:1,49,replace = T)

mydata <- as.data.frame(t(rbind(states,p_1,p_2,q_1,q_2)))

The chart layout below is what I would like to accomplish with one common legend.

Answer 1

You need to reformat your data so it is in long format with your variables identified by a key. Then you need to merge it with the spatial data. Then use facet_wrap(~ key) to create the four panels.

Try this:

library(dplyr)
library(tidyr)
us.dat %>%
  dplyr::left_join(
    mydata %>% 
      tidyr::gather(key, value, -states),
    by = c("region" = "states")
  ) %>%
  ggplot(aes(x=long, y=lat)) + 
  geom_polygon(aes(group=group, fill = value),
               colour = alpha("white", 1/2), 
               size = 0.2) + 
  theme_bw() + 
  theme(# legend.position = "none", 
        line = element_blank(),
        axis.text = element_blank(),
        axis.title = element_blank(),
        strip.background = element_blank(),
        panel.border = element_blank()
        ) + 
  coord_map("polyconic") +
  facet_wrap(~ key)

I added a few theme elements to get it to look similar to what you want. You will need to use a scale_fill_manual() to get the colors you want.