Why Pie Charts Suck But They Don’t Have To
The universal view of pie charts is that they suck. You throw one of these into your PowerPoint presentation and you better be prepared to be mocked mercilessly. But why are these charts so disregarded and hated? What did they ever do to make everyone hate them? Let’s explore how pie charts got a bad rep and why they are actually a very good chart.
Pie charts are meant to be used to illustrate numerical proportions. Each “slice” is proportional to the quantity it represents. Obviously called a pie chart because it looks like slices of pie. It had everything required to succeed but like most great things, people found a way to ruin it. The downfall of pie charts was ultimately user error.
Too many slices.
There is a limit to the number of slices you can have in a pie chart before it loses its context and meaning. That number is well below 20. I like to see the number of slices below 8, but as long as you can have context from the number of slices then it’s okay to have numerous slices.
Percentages not equal to 100.
It can cause confusion when you expect pie chart slices to equal 100 percent and it actually doesn’t. Your pie chart should represent portions of a whole, meaning all parts should add up to 100 percent. This allows you to use a pie chart without showing percentages because the assumption will always be that the slices add up to 100 percent.
Slices that look equal but are not.
Pie charts are great at showing a big discrepancy between a large portion and other smaller portions. What it is not great at showing is slices that are almost equal but not. When your values are similar in percentage then it can be hard to tell if slices are even, slightly larger, or slightly smaller. Slices like this makes your pie chart confusing and can lead viewers to believe items are even when they are not.
Used a pie chart when you should have used another chart type.
There are lots of options when deciding how to display your data: pie charts, line graphs, bar graphs, histograms, etc. It’s honestly extremely hard to pick the right chart type every time and opinions vary on the best way to display each type of data. You have to use your best judgement and ask yourself, “Does this best display the context of my data?” before deciding a pie chart is right for your data.
Now you know some of the pitfalls of pie charts. Even with this knowledge it can be hard to build a good, contextual pie chart. So you might be thinking, “What are pie charts good for then?” And to that I would say, “A few things!”
One large portion compared to smaller portions.
When you have one portion that far exceeds the other portions, this can be easily represented in a pie chart. Viewers can see, which much explanation, that one portion is larger than the others and therefore can make some assumptions on what you are presenting. It gives insight without needing a ton of context.
Values that aren’t equal.
If you are trying to show values that have equal or almost equal values, a pie chart is maybe not the right solution. If you have significantly different values and only a small number of slices then a pie chart can work.
Small number of slices.
I said this before but a small number of slices works best with a pie chart. I personally like anything below 20, even better if it is below 8. As long as you don’t lose count of slices when looking at it a pie chart works well.
Portions equal to 100 percent.
All your slices should add up to 100 percent. You need to show all portions, otherwise your pie chart doesn’t make sense and gives the wrong insight by viewing it. The assumption of the viewer will likely always be that slices add up to 100 percent.
You should find a use case where all these pieces of advice converge; small number of slices equal to 100 percent and values that are not equal. This is the perfect pie chart, in my opinion. Don’t be afraid to use a pie chart, they can be a good way to visualize your data. Use your best judgement and put those pie charts in your next presentation or IoT dashboard.
