Electoral GIS

[gfish]

There have been a lot of interesting maps floating around, looking at the election results in different ways. One of the interesting things was the emergence of the 'shades of purple' maps, looking at the popular vote in each state or county. Which is cool, and it's good to remember that no state is all one color. But it got me thinking -- which view is actually more correct? The first one implies a massive geographically based split in the nation. The second refutes that. But without a baseline, how are we to compare the two? I don't know what a critical geographically based split in the nation would look like rendered in shades of purple. The shades of purple view, while visually reassuring, could still indicate something horrible. How would we know? Luckily, there is an obvious comparison -- the Civil War! Almost by definition, there has been no time at which we were so divided. And, of course, we split along geographic lines. So what did the nation look like then, in shades of purple?

I had trouble finding popular votes by state. The best I found was this University of Virginia collection of maps, which unfortunately only gives popular vote percentage for the winner of each state. Yay that they made these maps available, bah that they don't provide the raw numbers.

I originally intended to use the 1860 election, but I had forgotten that 4 candidates got electoral votes that year. So I fast-forwarded 20 years and settled on the 1880 election. Only two real candidates, and it was very close in the popular vote. Politics were still strongly divided north/south. The data still isn't perfect -- I was only interested in Democrat versus Republican, and the percentages don't always add up to 100%. The Greenback party got about 3% of the vote, which is mostly visible in Virginia. It went 46% to the winner (Hancock, D). Obviously Garfield didn't get 54% of the vote, so I logged it as 44%. This isn't science, folks, just idle curiosity. I reversed the color scheme in my map to make comparison with the 2004 results easier.







This is hardly a rigorous or quantitative analysis, but I'd say we look about equally divided. What do you guys think?

Comments

[craigp.livejournal.com]

Just to be pedantic, according to this map there are 3 states that are in fact all red. http://www.usatoday.com/news/politicselections/vote2004/countymap.htm

[gfish.livejournal.com]

By county, sure. But the popular vote in the state wasn't 100% Bush.

[dymaxion.livejournal.com]

that's just who won each county, though — almost without exception, every red county on that map had Kerry voters in it, and vice versa.

[callmem.livejournal.com]

Now I want to see the county by county map for 1880. I like the county by county for the 2004 election because it points out all the enclaves of liberal thought scattered through otherwise red states. It also showed an interesting phenomena you can see from yard signs in my area of CA; the coastal area is squarely blue but when you cross over the foothills into ranching counties, it gets surprisingly red. I'm guessing you wouldn't see the same kind of pockets of differing opinion on the 1880 map as you do on the 2004. (which is why I'm resisting the urge to start a secession campaign a la Ecotopia...) (plus also I don't have nukes anyway)

[vixyish.livejournal.com]

Hey, I've been ready to secede ever since I first read that book. I think we'd have to relinquish southern CA altogether though, and move you folks up to northern CA or so.

[callmem.livejournal.com]

Sounds good to me. Just give me a heads up when all the paperwork goes through, 'kay? ;)

[gfish.livejournal.com]

Find the county level data and I'll gladly plot it.