Pie charts were cool in middle school when you needed to hear the word “pizza” to pay attention in math class. But when it comes to offering real insight using data visualization, do everyone a favor and avoid them at all costs!
If pie charts are so bad, then why are they everywhere? It baffles me. Pie Chart Users claim that pie charts are useful because they show how different parts make up a whole. But in most cases, they’re not even that good at it.
Here’s where pie charts fail us:
It’s hard to compare different data
The purpose of any chart is to make information and data more easily digestible – and, usually, to make it easy for people to compare several sets of data. But pie slices are not conducive to comparison. Sure, we can figure out that a gigantic pie slice is bigger than a tiny one, but what about comparing two similar-sized slices? It’s hopeless. Even if you labeled pie slices with percentages, it’s still not as easy as comparing the 2 data figures in a bar chart, where you’d be able to see exactly how much bigger one number is than the other.
The lesson: If two data points are similar in size, don’t use a pie chart.
You can’t measure data over time
To illustrate this, our engineers created a hack called “Pieify” that turns any InsightSquared bar chart into a pie chart. Here’s an example of a chart that shows Win Rate by month:
In this pie chart, each month is represented by a slice of the pie, and the size of each slice is determined by the percentage of overall opportunities that were won in that month.
Okay, so you can kind of tell that the slices get bigger with each month. But it is difficult to tell with pie slices how much bigger January 2014’s slice is than February 2013’s even though they are next to each other. Sure, one is huge and the other one is tiny, but how much bigger is the January 2014 slice? It’s much easier to tell when the data is visualized in a bar chart:
(Click on the picture for a closer look at our Win/Loss chart.)
This bar chart helps us see the timeline and size comparisons of won opportunities each month much more intuitively. In other words, it is way easier to read, understand and compare.
The lesson: If time is a variable in your data, don’t use a pie chart.
Too much information on a pie chart is really confusing
Take a look at the pie chart below, which is supposed to compare population of all 50 of our beloved United States.
What IS that? I get vertigo just looking at it.
The only takeaway I get from this is that California, Texas, New York, and Florida have the biggest populations – after that, everything gets fuzzy. Plus, the print is way, way too small, and there’s too much of it.
The lesson: If you want to compare more than 2 or 3 data points, don’t use a pie chart.
The only way pie charts can work
If you have to – HAVE to – use pie charts, make sure:
You use only 2-3 slices (as in Male/Female or Democrat/Republican/Independent)
There is a significant difference in slice size
You label the slices clearly
Don’t think of those tips as permission to use pie charts, though. They’re not that useful and they’re easy to mess up.
Hey – know what else is the worst? Spreadsheets.