Creating Maps With GIS Data in SAS 9.1.3, Part 2

By Patrick McLeod, Research and Statistical Support Services Consultant

In Part 1 of this series we looked at some of the basics of creating maps in SAS. In Part 2, we’re going to look at a couple of different map options in SAS that can help you create maps to add content and context to your data. I’ve already covered choropleth maps in Stata; in this column we will be revisiting choropleth maps and looking at prism maps, both created with PROC GMAP.

The data I’ve selected for this exercise is some data from a statewide race from the November 2008 election in Texas. This race was between incumbent Republican Justice Phil Johnson, Democratic Justice Linda Yañez from the 13^th Court of Appeals and Libertarian Drew Shirley, a lawyer from Austin. Rather than focus on statewide county-by-county returns, I wanted to create some maps that focused on three metropolitan areas and the collar counties surrounding them. I live and work in the county seat of Denton County and I am very interested in how the political dynamics of urban-suburban-exurban communities and the counties in which they reside. Rather than showing county-level returns for the entire state, I only reported data for the following counties (listed by metropolitan area): In North Texas, Collin, Dallas, Denton, Ellis, Johnson Kaufman, Rockwall and Tarrant; in Central Texas, Hays, Travis and Williamson; on the Gulf Coast, Brazoria, Fort Bend, Galveston, Harris, Montgomery and Waller.

Because I would like to create some measure of Republican intensity in these counties, I took the difference between the number of votes for Justice Johnson versus the number of votes for Justice Yañez and created a new variable labeled margin. Dallas, Harris and Travis counties all voted for Justice Yañez while the remaining counties all voted for Justice Johnson. The colors in the maps below reflect the scale of this margin variable, ranging from blue (Yañez) through yellow, orange and red (increasing margin of victory numbers for Justice Johnson).

Figure 1: Choropleth Map of Margin of Victory In Texas Supreme Court Place 8 for Selected Counties, 2008.

Figure 1 is a one dimensional county-level choropleth map of Texas that is projecting the margin variable I created for the Texas Supreme Court, Place 8 election from November 2008. This map nicely illustrates the hunch I followed when I began this project, namely that the Democratic counties of Dallas, Harris and Travis are surrounded by counties with significant to marginal suburban and exurban development that are, at least in the case of this race, varying degrees of more Republican than their urban neighbors.

Figure 2: Prism Map of Margin of Victory In Texas Supreme Court Place 8 for Selected Counties, 2008.

I like the idea of the prism map for this project because it shows both color and height to add an element of project to the map. I am a bit disappointed in some of the rendering artifacts along the lower Gulf Coast region of the map (the lines that extend down along the coastal areas) and I hope that PROC GMAP in SAS 9.2 might lead us to a cleaner-rendered image.

Happy Holidays from all of us in Academic Computing Services, Research and Statistical Support! We look forward to providing you with lots more reading material in 2009!