points.r

mid_range <- function(x) mean(range(x, na.rm = TRUE))
centres <- ddply(county_df, c("state", "county"), summarise, 
  lat = mid_range(lat), 
  long = mid_range(long)
)

bubbles <- merge(centres, unemp, by = c("state", "county"))
ggplot(bubbles, aes(long, lat)) +
  geom_polygon(aes(group = group), data = state_df, 
    colour = "white", fill = NA) +
  geom_point(aes(size = rate), alpha = 1/2) +
  scale_area(to = c(0.5, 3), breaks = c(5, 10, 20, 30))

ggplot(bubbles, aes(long, lat)) +
  geom_polygon(aes(group = group), data = state_df, 
    colour = "white", fill = NA) +
  geom_point(aes(color = rate_d)) +
  scale_colour_brewer(pal = "PuRd")

polygons.r

library(ggplot2)
library(maps)

# First (and most annoying) task - get matching state and county variables 
# for both datasets.  And unfortauntely it's not quite right, as you can
# see from the finish product - some counties are missing.
unemp <- read.csv("unemployment09.csv", header = F, stringsAsFactors = F)
names(unemp) <- c("id", "state_fips", "county_fips", "name", "year", 
  "?", "?", "?", "rate")
unemp$county <- tolower(gsub(" County, [A-Z]{2}", "", unemp$name))
unemp$state <- gsub("^.*([A-Z]{2}).*$", "\\1", unemp$name)

county_df <- map_data("county")
names(county_df) <- c("long", "lat", "group", "order", "state_name", "county")
county_df$state <- state.abb[match(county_df$state_name, tolower(state.name))]
county_df$state_name <- NULL

state_df <- map_data("state")

# Combine together 
choropleth <- merge(county_df, unemp, by = c("state", "county"))
choropleth <- choropleth[order(choropleth$order), ]
# Discretise rate to use with Brewer colour scheme - many options here
# choropleth$rate_d <- cut_number(choropleth$rate, 5)
# choropleth$rate_d <- cut_interval(choropleth$rate, 5)
# Nathan's choice is a little odd:
choropleth$rate_d <- cut(choropleth$rate, breaks = c(seq(0, 10, by = 2), 35))

# Once you have the data in the right format, recreating the plot is straight
# forward.

ggplot(choropleth, aes(long, lat, group = group)) +
  geom_polygon(aes(fill = rate_d), colour = alpha("white", 1/2), size = 0.2) + 
  geom_polygon(data = state_df, colour = "white", fill = NA) +
  scale_fill_brewer(pal = "PuRd")

# Takes a while to draw because ggplot2 not very efficient with large numbers
# of polygons :(

This was a very useful comment. Thanks!
@hadley: Thx for the code; I'm very new to maps and right at the moment don't find the time to dive in properly with a book, so working examples are a real time saver.

I had trouble running the codes. But after making following changes it worked.

library(scales) and replace pal = "PuRd" with palette = "PuRd" as suggested by shawneeunion above.

Thanks Professor Hadley, for the template solution. I get the following error, when I run finally the ggplot. Until then there is no error. Thanks for your help.

ggplot(choropleth, aes(long, lat, group = group)) +

• geom_polygon(aes(fill = rate_d), colour = alpha("white", 1/2), size = 0.2) + 
  

• geom_polygon(data = state_df, colour = "white", fill = NA) +
  

• scale_fill_brewer(pal = "PuRd")
  

Error in f(..., self = self) : attempt to apply non-function

R.Version()
$platform
[1] "x86_64-w64-mingw32"

$version.string
[1] "R version 3.2.3 (2015-12-10)"

On Windows, the slowness seems to be caused by alpha(). Changing this line

colour = alpha("white", 1/2)

to the following

colour = "gray85"

makes the plot draw almost instantaneously with only a slight change in the county border color.