Why use one number, when seven 3D graphs will do?

My first example come from the New York Times. This venerable publication, which in most respects adopts a sober, factual approach to the news, has for some reason become a bastion of what Edward Tufte has labeled “Chart-Junk.” One offender is Charles M. Blow, the “visual Op-Ed columnist” for the New York Times. Now Mr. Blow is a serious journalist, and the topics he deals with are interesting, important, and well described…in the text. But in his graphs Mr Blow seems to delight in the most obscure, gaudy, and expensive (in column inches) representations possible.

Consider a recent example from an Op-Ed piece of October 23, 2010 (http://www.nytimes.com/2010/10/23/opinion/23blow.html). Here is a picture that represents marijuana arrests of whites and blacks in several california cities.

OK. Quick glance. Pretty circles. Nice colors. What point jumps out at you? Exactly: nothing. This elaborate 3D graphic, spanning a total of seven separate pictures of small discs floating over big discs, with a nice shadow, serves to depict a very simple fact that is effectively obscured by the diffuse presentation.

Here is an alternative.

Yeah, I know. Not very sexy. But compact, legible, and to the point. The main conclusion jumps out at you: black arrests per 1000 population are vastly higher than white. A secondary conclusion – that some cities have much higher arrest rates, is also easily grasped, since we have arranged the cities in descending order of arrest rate. But since that was not really the point of the Times column, why plot it? Here is a second draft, focusing on the main point alone.

 

Here we show only the ratio of arrest rates – the main point of the article – as a function of the cities. Again we sort the results to show in which cities this disparity in arrests is largest. But actually, the variation over cities is not great, and also not a point of the article. So why show it? Thus we are lead to the somewhat dispiriting observation that there is really only one number of interest: the arrest rate of blacks versus whites averaged over all these cities: 5.71. This number certainly validates Mr. Blow’s assertion of “a grotesquely race-biased pattern of arrests,” but no graph is required to make the point.

Graphs are a wonderful tool to convey data in a manner that is immediate, intuitive, and compelling. But they can also be used to obscure.

Advertisements

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 )

Twitter picture

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

Facebook photo

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

Google+ photo

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

Connecting to %s

%d bloggers like this: