Visualizing TV Viewers - Sports and Politics

April 14, 2010
Posted by Jay Livingston

How do you turn data into a good graph? Of course you could ask flâneuse . But suppose you wanted to do it yourself.

Here are the results of a study on preferences in TV sports and in politics – 218,000 interviews conducted over a 13-month period. I’m not sure what the questions were that determined the Democratic and Republican index. The other variables, “Likelihood of voting” and being “very interested” in watching the sport on TV, are fairly straightforward.

The data in the table are sorted on the politics column (R-minus-D Index). PGA golf has the most Republican audience, WNBA the most Democratic.

(Click on the image for a larger view.)

How would you graph the data?

Here’s one possibility, found at dqydj (which stands for “Don’t quit your day job,” but you knew that already, didn’t you?).

(Click on the image for a slightly larger view.)

Blue bars represent political leaning – the difference between the GOP and Democratic indices. Green bars show likelihood of voting. Sports are listed on the x-axis.

I prefer this one, found here.

(Click on the image for a larger view.)

For more on creating visualizations, go to Many Eyes , which has a ton of data sets to play around with.

(Hat tip: Andrew Gelman)

No comments: