Following up with my previous blog titled "This Is Not a Data Visualization," I came across a nice example of what I am defining as the difference between data 'visualization' and data 'illustration.'
Randall Olson writes about the "United States of Amoeba," a visual representation of the cross-aisle network of the 113th United States Congress. This network diagram appeared in several news outlets, and it claims to demonstrate polarization in U.S. politics.
Randall nicely dissects this view by showing how a visualization becomes an illustration through selective reduction of data to fit the argument or viewpoint that the author wishes to advance.
An illustration is our reflection on reality; it is our filter that we use to communicate an idea. We use illustrations to express our own points of view. Caricatures are illustrations, in particular illustrations that highlight certain human features for the sake of humor or satire. Because of this light or gross distortion, we don't see caricatures on the front pages of objectively straight news outlets. We see caricatures in the opinion pages.
The problem that Randall appears to have uncovered here is a distortion (via a selective reduction) of data. While it's debatable whether this is a gross distortion, the distortion is nonetheless meant to advance an idea.
And here we have to think about the context:
Is the visual being used to further rhetoric, or rather exploration and discovery?Data illustrations (in the context of rhetoric) are based on reflections and filters of original data. Data visualizations (in the context of exploration and discovery) should be completely and purely representing the data.
I like how Randall shows us that we need to continue to look at visual representations of data with a critical eye.