Honest Maps

Dammit, BBC! Your reporting is usually so good. Then you go and use this map:

misleading map of US 2016 presidential election results

I never want to see that map again. Reporting election results by square mile is completely meaningless. It’d be like reporting the DOW and S&P 500 by number of trades made. It tells you something, but not what it’s claiming to tell you (and could be wildly misleading).

And the rest of the article is great. It immediately followed this map:

accurate map of US 2016 election results

Which is a great map! It shows us what it says it shows us. It’s easy to read, and not misleading.

Journalists: never use a true-geography map to report election results. Just don’t. Particularly for the US. (It might work OK for states or smaller countries, if they’re of more uniform population density, but that just means the amount of distortion is smaller, not that it’s gone.)


4 comments on “Honest Maps

  1. Dan Bongert says:

    Actually, that square map is pretty misleading as well. California (55 electoral votes) isn’t 2x the size of NY (29), and NY and TX (38) are pretty much the same size). It needs much better scaling.

    • woodelf says:

      Well, ok, it’s not perfect. It’s still a damn sight better.

    • woodelf says:

      Looking again, I think they’re scaled by area. I’ll have to do some measuring to see. It unfortunately looks like it’s not a linear scale—or at least the tiny states are oversized.

    • woodelf says:

      OK, I measured. The squares are scaled linearly (by area), within the margin of resolution. It looks like it’s about 240 square pixels per electoral vote.

      state votes side area area per vote
      WY 3 27 729 243
      NY 29 84 7056 243.310344827586
      CA 55 115 13225 240.454545454545
      WI 10 49 2401 240.1
      PA 20 68 4624 231.2
      FL 29 83 6889 237.551724137931
      WV 5 34 1156 231.2

      Tufte’s best practices for this sort of visualization is to scale by area, not by linear dimension, because for most people the visceral estimation is closer to the actual magnitude of change. There is an argument that you need to split the difference—maybe scale by a linear factor of √3—to better approximate how we perceive things, but last I checked it’s unproven.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s