Criticism

Terrible charts seem to be in season. Rarely have I come across so many incredibly bad charts in such a short time: information graphics that don't actually depict data, distorted representations, useless color schemes, and the worst pie chart ever.

For some time now there has been some discussion about finding a new terminology for the Data Visualization field. The intention is to find names that reflects the two main different directions that are seen today on Dataviz: visual data analysis and data driven aesthetics images. The concern is that Data Visualization might lose its ‘serious’ role as an analytic tool by including works that, without doubt, praise aesthetics before clarity.

I recently criticized Ben Fry's visualization of health care cost data from GE and claimed that I knew how to do it better. While my analysis may not be as pretty and flashy as Fry's, it provides actual insight into the data. It also reveals an interesting issue: the data is really dull. So dull, in fact, that a visualization was needed to cover up that fact.

General Electric recently commissioned Ben Fry and Seed Media to visualize health data to communicate the costs of different kinds of diseases to the public. The result is pretty and colorful, but of little value if you actually want to learn something.

Coming from the academic and computer science side of visualization, I always assumed that it would be self-evident to anybody that visualization is first and foremost useful, and only happens to also produce nice pictures. Alas, this is not actually the case. To most people, visualization means pretty pictures first, and maybe also a fact or two. We have to fight that or risk the trivialization and marginalization of visualization as an analytic tool.

Of all the sins committed against visualization on the Internet, the Periodic Table of Visualization Methods stands out as the most egregious. Its collection of actual visualization methods, structural diagrams, and feel-good business bullshit does not fit a structure that was devised to understand the world – and that is actually a very effective visualization in itself.

One of the most common mistakes in chart design is to scale an area by two sides at the same time, producing a quadratic effect for a linear change. That overstates the larger numbers and produces a badly skewed chart. A little care and some basic high-school math can help avoid the problem.