Graphs Worth a Thousand Words

The background:

This article from the New York Times, explaining how the Obama campaign is trying to get Virginia to swing for a Democratic presidential candidate (which would be the first time in 44 years it would happen). The 2004 wasn’t close (Bush got 262K more votes than Kerry), but Obama would like to focus on areas where Bush won by a smaller margin, which hopefully indicates that voters living there would be more willing to vote for a Democrat.

The need:

The New York Times wanted to show how voters were distributed across the state, especially which areas were more heavily red or blue. Most charts turn states, counties or cities red or blue depending on whether they are trending Democrat or Republican. That’s important, but oftentimes, more fine-grained detail is necessary.

The solution:

This graph could have been done with shades or red or blue. Deeper reds would indicate a larger Bush margin, deeper blue for Kerry, and gray or purple could indicate a middle ground. However, the entire county would be shaded that color, so color saturation would indicate size of the margin. This is graspable by the mind, but not a perfect correlation (greater color saturation = larger Bush or Kerry margin).

Instead, the New York Times chose to indicate bigger margins with bigger circles covering the area of interest. What is interesting is that this still gives the information of which areas are deeply red and blue and which areas are split (split votes indicate no margin for one candidate or another, leading to very small circles). However, this method correlates much better with the underlying concept (larger circles = larger Bush or Kerry margin). A big circle intrinsically means “Wow, Bush really dominated in this county” whereas a really deep red would likely not obviously indicate such a relationship.

It’s a small change in how these numbers are graphed, but it makes the data more easily understood, which is no easy task considering how much data there is out there that must be sifted through.