Why Pie Charts are Evil 

Pie charts were probably the second or third graph type your math teacher introduced to you. Maybe they even forced a protractor into your hand so you could get all the angles and percent ratios correct. 

Today, pie, donut, and gauge charts have been exiled from the data analytics and visualization community. Data experts rave about how they are “inadequate,” “bad,” or even “evil.” But why do pie charts have such a bad reputation these days? Let’s find out.

Pie charts take up too much space

Whenever visualizing data, you have limited space to work with. Either in the confines of a PowerPoint slide or as one step in your data story, you have to treat that precious digital real estate with care and fill it with the most value for the viewer/reader. 

The problem with Pie, Donut, and Gauge charts is that they use more space than alternatives that provide the same value (e.g., a bar chart). This is due to two reasons: the nature of their shape and how they depict information. 

• Pie, Donut, and Gauge charts use area to give value to a single category, whereas bar charts just use distance. 
• These charts are circular, which creates unnecessary space around the chart object.   

Pie charts aren’t as readable as other chart types

When looking at a pie, donut, or gauge chart, it’s difficult to tell how much larger one portion is than another. This is because the human eye and mind are much better at comparing distance instead of area. After all, you only have to assess one dimension instead of two.

For example, if I asked you “how much bigger is ‘Pinot Grigio’ than ‘Tempranillo’ on the chart below, it would be challenging to give an exact answer. 

Source

However, with a bar chart, you might have an easier time finding your answer: ‘Pinot Grigio’ is about 3X larger than ‘Tempranillo.’

Source

People also have trouble with their perception of angles. It’s easier for people to judge the magnitude of a slice when it has nice round 90° angles like 0, 25, 50, 75, and 100%. However, humans can’t easily perceive the differences in angles that fall outside those numbers, and, with no scale, it becomes even more problematic. 

We often overestimate obtuse angles and underestimate acute ones. Even when you have easy-to-read values (like 90°, 180°, 360°), something as simple as rotating the chart can still throw off your perception. 

As you add more segments, this problem gets worse. Pie charts also have problems fitting labels on the graph when there are too many variables, forcing the use of legends and wasting the viewer's time switching their eyes back and forth between the legend and the chart. When the slices get too small, people like to add labels and numeric values to each slice. However, if you’re adding in names and numbers, you might as well be looking at a table. 

Source

Pie Charts can’t convey more information without becoming overly complex

Pie charts don’t allow as wide a range of options for additional data points as bar charts and other types. There are several ways to enrich bar charts with additional data points that don’t make the graph look too crowded or hard to read. You can see this in the bullet chart below, which adds additional values to display targets and ranges. 

Source

However, adding these elements to a pie chart makes it harder to read. The same can be said for 3D renders of pie charts that don’t add additional information. Instead, they will just add complexity and require you to take another dimension into consideration. They will also distort the images, as whichever slice is farther away will look smaller.

Source

Source

Best Alternatives to Pie Charts

In 90% of cases, you can use bar charts instead of pies. They can display almost any information a pie chart can, and they’re easier to read. The only question is, what type of bar chart should you use? Remember that no matter what kind of chart you choose, your decision should be based on: 

• Your use case
• The data you want to chart
• The space you have available 

Many people think that percentages are best displayed on pie charts. However, bar charts are still easier to interpret for this type of data because you only have to assess one dimension. If space is limited and you only have a few values, a stacked chart can also be helpful.

On the other hand, a classic bar chart is an excellent choice when you have a lot of values and want your viewers to make a comparison quickly. 

Source

Some people choose pie charts just to add variety to their presentations. However, using the correct tooling is far more important. For data visualization, the proper tooling is always whatever is easiest to read and interpret.  Better to be understandable than exciting. If anything, having the same type of charts makes your reporting more consistent.

What are pie charts used for? 

While we do find pie charts to be inferior, no rule is written in stone. Pie, donut, and gauge charts can still fulfill a purpose. One example is the use of gauges in cars. 

Some reasons you might choose pie charts are:

• When there is only one slice of the pie, it’s easier to show its’ share of the whole value. 
• You have a square (not rectangular) space for visualization
• Users are used to this chart from past visualizations
• It’s too complicated to display the information as a bar graph (e.g., analog measuring devices).
• Data is in %, and only two values are displayed.
  • This allows the user to see data and percentages, and they can quickly reference the dimensions based on 50% 

Tips for using pie charts: 

• Order sections from smallest to largest starting at the 12 o’clock position
• Don’t have more than five sections
• Don’t use them to display change over time
• Don’t compare pie charts to one another
• Ensure high data quality before building the chart

Data visualization is more than charts

Charts are an excellent means for communicating the message behind your data. By understanding the different types and when to use them you can build compelling visualizations that always get your point across.