(This article was originally published at Junk Charts, and syndicated at StatsBlogs.)
This WSJ graphic gives me a reason to talk about the self-sufficiency test: go ahead, and block out the data labels on the chart, you are left with concentric circles but no way to learn anything from the chart, not the absolute dollar values, nor the relative dollar values. In other words, the only way to read this chart is to look at the data labels.
The online article does not include the graphic. It's an article talking about Neil Armstrong's death. Here's the same data using bar charts:
The chart would be much improved if a longer time series is included giving us values for each year. It's pretty clear that this data is subject to sudden jumps (e.g. Armstrong's death) and so picking arbitrary years will likely cause is to miss important events.
***
Circles are also subject to various types of optical illusion. Before you use bubble plots, give the following a look:
Can we judge the size of circles in relation to other circles? (credit)
Can we judge the relative distance between circles? (credit)
Can we judge the relative sizes of circles within circles? (credit)
Please comment on the article here: Junk Charts
