Choropleth Map R Challenge

FlowingData today has an article on how to make a choropleth map (a map with regions shaded by values of a measurement variable) using free tools. The goal is to recreate this map:

The free tool FlowingData used is Python, and the whole process takes 14 steps and involves finding an SVG file with the county-level map and editing the SVG file with a text editor (automating part of that process with an XML tool in Python). At the end of the day the result is an SVG file -- high-resolution, yes, but not compatible with all browsers.

Surely this could be done more easily in R? You can create a county-level map with the maps library and the map("county") command (or you can use this method with the sp package), and the data are available as a CSV file which you can read in with read.csv. RColorBrewer should even generate the same color scheme. So I challenge the R hackers out there - let's show 'em how it's done in R.

FlowingData: How to Make a US County Thematic Map Using Free Tools (via @marcoscan)

Comments

[Shane]

Why don't you post that as question on stackoverflow? I was thinking the same thing when I read that blog entry on FlowingData this morning...

[Hadley]

Here's my attempt- http://gist.github.com/233134. It would be a lot more succinct with a better data source for the counties. Any suggestions?

[Barry Rowlingson]

Took me about twenty minutes - I found a shapefile with FIPS codes, read in with readOGR from package:rgdal, then read in the unemployment data with read.table, matched up the FIPS codes (two missing), then converted things to a factor to plot with the same colour scheme, used spplot for the graphic.

Python's BeautifulSoup of course is made for hacking XML files. I'd actually like to see him do it properly in Python using a shapefile and the various python map packages.

Neither gadm:
http://www.gadm.org/country

nor maplibrary:
http://www.maplibrary.org/stacks/gadm/USA/index.php

seemed to have FIPS codes in their maps of US counties.

[Jason H]

Here is mine.

http://gist.github.com/233257

It craps out a bit on a few counties since the maps(county) data has some strange names. Fun exercise though.

[J]

Love the challenge, David - here's my take:
http://www.thisisthegreenroom.com/2009/choropleths-in-r/

[A.I.]

This might help:

Figure 13.10
http://lmdvr.r-forge.r-project.org/figures/figures.html

[Eduardo L]

And mine, using base graphics and census data.

http://gist.github.com/233956

http://files.eduardoleoni.com/map20091113.png

[Pat]

Good stuff...I used this as a springboard to create choropleth maps for all economic and demographic data in the u.s. census at StatJump

e.g. Check out the distribution of Italian population by county.

[clementine]

I'm trying to do the same thing but for only one state. I tried getting the shape file and using readOGR but get an error message. Any idea how I could go about it? cheers

[LA Food Truck Map]

Is the data for this available there still?