The purpose of graphics for data visualization is communication, not decoration, which is often forgotten in publications as well as on company performance dashboards. A case in point is the chart on yesterday’s cover of the New York Times. It shows that solar energy currently accounts for more than twice as many jobs as coal. It also shows the numbers of jobs in different sectors and uses a color code to mark some as based on fossil fuels versus renewable and low-emission technologies.
Until recently, most publications would have used a pie chart. Now, graphic artists have found a way to square the pie chart into yet another style that will most likely trickle down to slideware and office walls, in spite of a low data-to-ink ratio and the use of two-dimensional shapes to display one-dimensional data.
If we agree that a picture is worth 1,000 words, then you don’t need a picture if your information fits in 50 words. The numbers in the chart are from the 2017 US Energy and Employment Report, Table 1, p. 28, which I would summarize as follows:
It is, perhaps, less flashy but more informative. The emphasis on Solar vs. Coal is given by the red frame. As in the New York Times picture, the fill color indicates the type of source. Unlike the New York Times, however, I used a special color for Nuclear. Uranium sure isn’t a fossil fuel, so it doesn’t belong with oil, gas, and coal, but that doesn’t make it renewable. As for Nuclear being a low-emission technology, the inhabitants of Chernobyl and Fukushima might disagree.
The other issue with the New York Times chart is that it uses 2D objects — rectangles — to display 1D data — numbers of employees. If you absolutely want to show this data graphically, you could do it with a horizontal bar chart, which would make the relative sizes of the sectors more visually clear than rectangles. With rectangles, for example, it is not visually obvious that the Other category employs more people than Wind; on the bar chart, on the other hand, it is, because the bar for Other clears the 100,000 mark while the bar for Wind doesn’t.
In slideware and business reports, the most common graphic for this kind of data is the pie chart, where the one-dimensional Employment variable is represented a the one-dimensional wedge angle, but it is still not as information-rich nor more visually clear than the table or the horizontal bar chart, as you can see in the visually indistinguishable wedge sizes for Wind and Other.
In conclusion, for the purpose of communication, yesterday’s New York Times squared pie chart is the worst. and the traditional pie chart is not much better. The table, and the horizontal bar chart work best.