In the world of data visualization, lore has it that famed information design guru Edward Tufte said that people who use pie charts should be viewed with the same skepticism as those who mix up “its” and “it’s.”
While Tufte’s comparison may be slightly unfair, there’s a point to be made that underlies his contempt for the infamous pie chart: pie charts are almost always the worst option for visualizing data. And…those who don’t realize this are, essentially, poor data communicators.
So, what’s the problem with pie charts?
At best, when pie charts are used correctly (like, according to the basic rules, where they should only be used to compare parts of a whole, in percentages that add up to 100%), they don’t do an effective job at storytelling or emphasizing key messages. At worse, when used incorrectly, pie charts make no sense at all and only make the person who created it look totally oblivious.
When deciding between bar charts, area graphs, line graphs, dot plots, histograms, or any other data visualization, the fundamental truth is that there is almost always a better chart to use than a pie chart.
Check out my (somewhat facetious, clearly) decision tree to help you determine if a pie chart is the best chart for your data (and click to enlarge):
If you ask pretty much any data visualization expert, they’ll tell you the same thing: pie charts are akin to Comic Sans on a professional résumé. Gross.
So, When Should I Use a Pie Chart?
If the above chart is any indication, pretty much never. Some will go so far as to say that you should never, ever use a pie chart. Why? Because they’re difficult for the human eye to make comparisons, they muddy and manipulate data, and they’re usually the least effective out of an array of options for highlighting your key message.
That said, I always hate to go so far as to say you should never use pie charts. There’s likely always a time and a place. If you do choose to use a pie chart, there are five basic rules for correct use:
- Your data needs to be summarized as parts of a whole (like the population of a specific country);
- Your data needs to be able to be represented in percentages that, when added up, amount to 100%;
- Your variables should be drastically different (do not have variables of 16%, 14%, 19%, 13%, 15%, and 17%, for example);
- You should only be comparing three or four variables–never more than five–in a single chart.
- You should never use two pie charts side-by-side to make comparisons between data sets.
Of course, there are also some other visual rules for using pie charts:
- Never create 3D pie charts;
- Never make them “exploded”;
- Never use conflicting colors;
- Don’t put too much text or numbers in the slices.
What Should I Use Instead of a Pie Chart?
Unfortunately, there’s not a real clear answer to that since there are, literally, scores of chart and graphs to choose from. But, even at the most basic level, bar charts usually work better than pie charts. Determine what your data set is and what the key message you want to share is. See my page on data visualizations to see some of the options you have at your disposal.