In Defense of the Lowly Pie Chart

The poor pie chart has been much maligned by data visualization guru Edward Tufte and more recently in this Voice of Vovici blog post.

Really, what has that poor pie chart done that makes it so abhorrent to folks? Yes, overuse and misuse are the main culprits, but that can be said of many charting options.

 

The beauty of a pie chart is that each ‘slice’ is instantly recognizable as a proportion of a 100% total. The same cannot be said for the bar chart…without the scale below, one would have to rely on grade school arithmetic to quickly add up the sections. Is the pie chart glamorous? Nope, but it is quick and easy for readers to understand the proportions you are reporting about.

 

 

I admit that I was nice and put the actual data labels (aka: percentages) in both charts, but just envision them without and see how easier the pie chart is to read…

Now where the pie chart fails miserably is when comparing more than one data set. In the example below we have three products that respondents are either currently offering, planning to offer, or have no plans to offer. It’s much easier to line up three bar charts (ensuring the 100% scale is there!) than try to compare the three pies.

 

So, think about whether your reader needs a quick visual of a single data set…and maybe the lowly pie chart is just the ticket.