This week our class read the first two chapters of Alberto Cairo’s new book, How Charts Lie. As a scientist-in-training myself, I would like to think that I am pretty good at reading and interpreting charts, but a little tune-up never hurts. I had actually read on NPR about the visual that Cairo opens with, the infamous map of the territory that Trump had conquered but I never paid attention the the words that were associated with the map. I have to admit that I didn’t recognize the fallacy in the map how When we are taught how to read graphs and charts in science classes, we are actually taught to ignore most of the words and try to interpret the charts a priori. The map displayed below isn’t by itself a chart that lies, but when an annotation layer is added indicating that this chart represents “Citizens for Trump,” it becomes a misleading graphic. Image of Trump District Map

Another important point that Cairo points out is that encoding (representing data as a visual elemenet) the magnitude of a feature in 3-D space is extremely challenging. In my reading of scientific literature, I have seen very very few examples of 3-D line charts or pie charts. One exception to this rule is that sometimes high dimentional data sets will be represented on a three dimentional scatterplot with each axis representing a principle component. Given that Cairo gives so many examples of charts misleading the reader, I am wondering if there is ever a useful way to use 3-D when trying to communicate information.

3-D Chart

The above chart is an example that I found with a quick google search. The bar graph forces you to occupy a perspective and therefore leads you to think that the bars closer to you are larger. This reminds me of the challenge that cartographers face in projecting the globe onto a 2-D surface. Something will be distorted when taking 3-D image and putting it into 2-D. But from Cairo’s argument, it seems like the converse applies as well.