by Andrea Hutson
You don’t need to search very hard on the web to find people that cannot stand pie charts. In fact, one of the first posts I found when searching for the topic was The Five Stages of Grief Over The Death of Pie Charts.
(Great read, by the way.)
But I think, as with many hated data visualizations, pie charts have their place, and today I’d like to share some things I’ve learned using Covid-19 as a timely example. (If you’re reading this way in the future, Covid-19 or Coronavirus was a nasty lung-infecting virus that swept the world in 2020, devastating countries and entire economies.)
I’m going to give a few examples of best practices for making pie charts useful. Much of what I’ve learned came from Ann Emery’s amazing course, Great Graphs. At publication time, Great Graphs wasn’t open, but I highly recommend following Ann and trying out her other courses. She’s at Depict Data Studio.
Do’s and Don’ts for Pie Charts
Don’t: Use too many categories.
Top 10 Countries with Most Confirmed Cases as of March 8, 2020
This one is a rainbow of awful. What’s wrong with this chart? Several things, but the most important is that it has far too many categories. There are 11 in total, and all but the top 5 are almost impossible to see. Go ahead, try to locate Switzerland.
I’ll wait.
This chart forces you to use the legend to understand what’s going on, and in the end, the legend becomes the only thing that actually gives you relevant information. So what’s the point of a pie chart here? I would recommend converting this instead to a bar chart if you want all of the information. But first think – do my readers need all this information? Why am I presenting data about 10 countries instead of 7, 11, or 200?
If you can lower the number of categories, I give you full permission to If use a pie chart.
Do: Keep the number of categories to 5 or fewer
Top 10 Countries with Most Confirmed Cases as of March 8, 2020
This is the same data, but with the number of categories kept to 5. Now we can actually interpret the pie chart itself, without relying on the legend.
Honestly, 3 or 4 categories is ideal. Keep in mind that if some of your categories are really tiny, unless that is a fundamental part of your message, it is likely better to use another type of visualization.
Do: Include the percentages and labels inside or next to your pie chart
Second, include your percentages and your data labels IN the pie chart. Ditch the legend. A good rule of thumb: if, in the end, you need a legend to interpret your pie chart, consider using another form of data visualization.
Do: Let the title tell the story
This is a point I try to make with every graph, and it’s so important. When you make a data visualization, you are telling your reader a (hopefully compelling) visual story. You are editing what you show to make a point. Why not explain the point so that readers don’t have to figure it out (or misinterpret the findings?)
The Vast Majority of COVID-19 Cases in the US are from an Unknown Origin
The pie chart above is pretty clear on its story, but having the title spells it out for the reader. Sometimes the mere act of typing the title will also help you to make your graph more clear to make sure the message is consistent.
Do: Get creative
I used an exploding pie chart in the above visualization. Data visualization experts will tell you that you should never use exploding pie charts, but don’t listen to those spoilsports. I used one here because I tend to think of viruses and the immune system as a Pac Man-type system. I think it works. Anything you can do to keep your reader interested is always a good idea.
(By the way, it is absolutely maddening that as of time of publication, it’s almost impossible to get tested for the virus in Central Texas and other places in the U.S. unless you have been to a place with a high rate of outbreak or have had close contact with a confirmed case. As you can see, this practice makes no sense whatsoever.)
In conclusion: don’t be afraid of pie charts, but don’t make terrible pie charts either. With just a few easy rules of thumb, you can make great ones. Want to learn how? Contact me to enroll in my 2-hour in-person class, Telling Stories With Data , where I walk you through the principles of good graph design in EXCEL. Or take Ann Emery’s course when it’s available. It’s fabulous.