America's true colours: purple and white

[madd0ct0r]

http://www.fastcodesign.com/1671268/inf ... =twitter#1



every 100 votes is a dot randomly placed in the county in which they were cast.

Every major city seems to be blue downtown and wrapped in red. no big surprise, but a great presentation of 'polarised america'

[LadyTevar]

What's the big areas of white?

[Grumman]

What's the big areas of white?

Places where people didn't vote very much, either because of low population density or low turnout.

[xthetenth]

You can actually make out that there's a faint speckling of dots there, but there isn't the thick saturated color of the more populous areas.

[madd0ct0r]

What's the big areas of white?

Empty space. only one dot for each 100 votes. Wyoming hasn't that many people.

[TimothyC]

Why isn't the placement by precinct? That data should be available in the vast majority of places.

By placing the dots randomly, counties with both significant republican and democrat vote totals in geographically distinct regions are diluted into a purple mess.

[Sea Skimmer]

If you go to the site you can get a version that lets you magnify it and see the dots in the sparser areas. Still kind of useless given the random spacing within counties.

[Simon_Jester]

I think the randomization is part of the point: sure, Democrats and Republicans are spatially separated, but by and large they're not that separated. It's easy to overlook that in a purely binary blue/red map.

[Grandmaster Jogurt]

At some point it seems to fill up the counties, losing data; there's no way New York or Chicago are equally for Obama and Romney, but they're middle purple.

[Tsyroc]

It's a much better map to look at than one that shows blue and red states. It makes the country look much more strongly divided. Even the ones that color counties make things look worse than this one.

[Grandmaster Jogurt]

An even better one is something like this, where it actually shows the margin of victory, and the size of the county corresponds to the population. Sure, it could go by precinct, too, but it's better than the usual.

[Simon_Jester]

That one is that it distorts the spatial relationships. I can get a pretty good idea of what blobs of color match what parts of the country, but it makes it harder to map the data in the image to a real life concept.

[Col. Crackpot]

An even better one is something like this, where it actually shows the margin of victory, and the size of the county corresponds to the population. Sure, it could go by precinct, too, but it's better than the usual.

Presentation is as important as accuracy. That looks like a 2 year old's finger painting.

[D.Turtle]

For 2012:


The first one simply shades the county by vote share, the second one resizes the counties by population. It distorts the shape, but the first one makes large, but empty counties distort the actual results too much.
I also like this one, which has the population as the height of the county:

Link to the WebGL model (VERY LARGE)

[Keevan_Colton]

It is rather telling that not a single one of the large spikes are red.

[Eldy]

I think the best visual representation I've seen is the county map that represents population by the shading. You can clearly see the urban/suburban rings as well as the vast rural areas.

[Simon_Jester]

The thing that troubles me about maps like this large one at bottom (I can't really tell if this particular map does it)...

The color scaling isn't given. I suspect it exaggerates the visual difference between, say, 40% Democrat and 60% Democrat, in a way the 'purple map' does not. But I can't prove that.

It is rather telling that not a single one of the large spikes are red.

If you live in a densely populated area, odds are you vote Democrat. Some of us may be disappointed by the fact, but I don't recall hearing that Americans are supposed to only get a vote if they live in densely populated areas. A thousand square miles with ten people per square mile get as many votes as ten square miles with a thousand per square mile.