Roll your own Red State/Blue State map with Mathematica

Kathryn sez, "Wolfram Research has posted instructions on how to import polling data into Mathematica 6 and how to use an example from the Wolfram Demonstration Project to create your own Red State/Blue State map. In principle, this allows Mathematica users to introduce their own assumptions into the mathematics." Analyzing U.S. 2008 Elections with Mathematica (Thanks, Kathryn!)

[#1 posted by Eduardo Padoan , September 24, 2008 7:52 PM]

I remember that professor Andrew S. Tanenbaum (ask any compsci student who he is :) was responsible by a site dedicated to keep trank of the latest US presidential election (electoral-vote.com). I wonder what techinic he use to draw the map.

[#2 posted by Trvth , September 24, 2008 8:29 PM]

This looks like a good tool. Just how accurate have these polls been in the past? I've never been polled - have any of you? Were the poll choices misleading at all? I'd think it would be pretty easy to manipulate poll results by wording the poll questions in a confusing manner.

[#3 posted by coaxial , September 24, 2008 10:45 PM]

@2: Thats why the actual results (not the soundbite on the news) actually contains the questionare. Examine it yourself. It's on whatever polling firm's website you want.

[#4 posted by coaxial , September 24, 2008 10:47 PM]

@1: He simply grabs the most recent poll, which in turn causes some strange fluctuations in battleground states. It's not exactly the best system. RealClearPolitics weights the polls, and describes how they do it.

[#5 posted by jfrancis , September 25, 2008 3:25 AM]

Not fair.

They need to dull that red down until it has the same luminosity as the blue. It dominates the viewer's perceptions otherwise.

http://www.digitalartform.com/archives/2004/11/red_vs_blue_one.html

[#6 posted by escowles , September 25, 2008 5:26 AM]

@4: You're wrong about the electoral-vote.com algorithm. He's got a page describing the algorithm here: http://electoral-vote.com/evp2008/Info/map-algorithm.html

But basically they use the most recent poll, and any within a week before it, averaged equally.

[#7 posted by Drew from Zhrodague , September 25, 2008 5:30 AM]

You could also use something more geared towards making maps, like Mapserver ( http://mapserver.gis.umn.edu ). There is a bit of learning curve, but once you get it setup -- it works quite well. Then you'll need Census data, of course...