Stop using those ugly Pie Charts!
A “slice” of me dies every time someone asks me to include a pie chart on one of my reports or dashboards. Although the famous Pie Chart has been around for centuries, it’s nothing more than a silly sliced round shape. Allow me to explain…
The problem with pie charts is that the human eye is not very good judging angles. We overestimate obtuse angles, and underestimate acute ones. Since the point of a chart is to help readers understand the relative sizes of a dataset, it seems a bit paradoxical to say the least, don’t you think?
Humans are a lot better at judging lengths of bars, than angles. William Cleveland was not the first to “attack” pie charts, but he presented a scientific paper in 1984 where he found that the length judgments for the bar chart were 1.96 times more accurate than the angle judgments in a pie chart.
So why do we keep seeing these infamous charts?! I came up with four reasons to justify it:
- Data visualization is a relatively new science and not many people look at it with a scientific mindset. They fill their reports with pie charts, colors, gridlines, and yes… sometimes even 3D charts (that’s a whole different post/rant)! They put the artistic effect before readability.
- Circles are more pleasant to our brains. An interesting study from John Hopkins University shows that the human eye prefers circles to sharp angles. It suggests that our brain associates sharp objects and angles to danger, and round objects and curves to health and life. You can choose to be a skeptic about it, but they do have scientific evidence that supports this theory!
- Excel has pie charts in its chart library. And they will probably be there until the end of time, enabling anyone with little to no skills to perpetuate the pie chart on their own “reports”.
- Proper data visualization is not taught in school. Perhaps with the rise of Data Science we will see some change in this, and it gets incorporated somewhere in the Math courses.
Quick test. Look at the image above and pick the one in which the “No” slice is bigger.
If you have been reading this article, you probably guessed they are actually the same exact size. If you picked the right one, thank you for proving my point! As a bonus reason for the survival of pie charts, we have to consider the deliberate deception by the chart creator. It’s not necessarily lying. It’s just representing the same truth, but in a different “perspective”. Literally! So… in a way, pie charts are the Donald Trump of Statistics!
Now that we got this clear, you may wonder if there is any situation where a pie chart is actually useful. To be honest, there is only one situation where I might consider using them. If you are presenting data with only two possible variables (like Yes/No) and you visually emphasize one of them (for instance graying out the “No” slice and leaving the “Yes” in color). But even then you will have to add a numeric label. You can also have a third variable to determine the size of the pie (useful for presenting multiple pie charts), but that’s not going to work every time and it’s not a very common visualization.
In short, you should seriously reconsider using pie charts in your dashboards, and leave them only to simple infographics or less complex reports. Even if your audience does not thank you directly, at least your presentations will be interrupted a lot less by questions like “is that yellow slice bigger or smaller than the red?”.
