Tuesday, November 09, 2004
Alternate views of the electoral map
I love interesting ways of displaying complex data, especially maps. With the US presidential election results in, lots of interesting maps have been created.
I think we're all pretty familiar with the classic Red vs. Blue map, with states that voted for Bush being colored red, and states that voted for Kerry being colored blue:
Here is another view, where one box represents one electoral vote (from the New York Times):
The state-by-state view is a bit simplistic, of course. Here are the results by county (From USA Today):
That's a lot of red, but this too should be weighted by population to get an accurate view, and there are a number of different ways people have done this. This is the New York Times's method: circles that are proportional to a county's population.
This 3D view by ESRI (the company that makes some of the election analysis tools used by the networks) is a bit clearer. A county's height is proportional to its population (via The Big Picture)
But the most interesting view is from Michael Gastner, Cosma Shalizi, and Mark Newman at the University of Michigan, where they have used an algorithm to distort the shape of the county-by-county map to reflect population differences. Here is the fascinating result (via Crooked Timber):
They also did it on the state-by-state results, which is a bit easier to get your head around:
There are those who say that this red and blue view is a little too black and white. A state/county where a given candidate got 50.000001% of the vote is colored the same as a state/county where the candidate got 90% of the vote. Robert Vanderbei has put together a map based on this idea, and the result shows that there's not really a "Red America" and a "Blue America", but more of a "Purple America":
Gastner, Shalizi, Newman applied their algorithm to Vanderbei's map, and came up with this, my favorite map of the lot:
They also have their images in wallpaper size
This is a bit anticlimactic after that one, but it's an interesting map with different data, showing how 18-29 year olds voted. It's based on exit poll data, so it is of course skewed towards Kerry, but it's interesting nonetheless (via The Big Picture):
I think we're all pretty familiar with the classic Red vs. Blue map, with states that voted for Bush being colored red, and states that voted for Kerry being colored blue:
Here is another view, where one box represents one electoral vote (from the New York Times):
The state-by-state view is a bit simplistic, of course. Here are the results by county (From USA Today):
That's a lot of red, but this too should be weighted by population to get an accurate view, and there are a number of different ways people have done this. This is the New York Times's method: circles that are proportional to a county's population.
This 3D view by ESRI (the company that makes some of the election analysis tools used by the networks) is a bit clearer. A county's height is proportional to its population (via The Big Picture)
But the most interesting view is from Michael Gastner, Cosma Shalizi, and Mark Newman at the University of Michigan, where they have used an algorithm to distort the shape of the county-by-county map to reflect population differences. Here is the fascinating result (via Crooked Timber):
They also did it on the state-by-state results, which is a bit easier to get your head around:
There are those who say that this red and blue view is a little too black and white. A state/county where a given candidate got 50.000001% of the vote is colored the same as a state/county where the candidate got 90% of the vote. Robert Vanderbei has put together a map based on this idea, and the result shows that there's not really a "Red America" and a "Blue America", but more of a "Purple America":
Gastner, Shalizi, Newman applied their algorithm to Vanderbei's map, and came up with this, my favorite map of the lot:
They also have their images in wallpaper size
This is a bit anticlimactic after that one, but it's an interesting map with different data, showing how 18-29 year olds voted. It's based on exit poll data, so it is of course skewed towards Kerry, but it's interesting nonetheless (via The Big Picture):
<< Home
