Hi there! In this 'Time to Viz', the infamous pie chart gets its turn. Even though a lot of visualizers hate or even despise the pie chart, in some occasions it can be a good visual. In many other occasion, however, a different visual is better. Pie charts are often used to expand the visual diversity of the dashboard, or that the perception lives that no other chart could represent the data. When looking at a pie chart, it always represents data in form of slices, where each slice represents the percentage of the total value.
A bad pie
A wider community of visual data artists would agree that the pie chart should be banned from most tools and dashboard and the reason is pretty simple: they are not easy to interpret.
Let's take a look at the following pie chart:
Feel uncomfortable, right? No way any user could actually get insights out of this chart. Then what is a good pie chart visualization you might ask? Well, the best way to come up with a good pie chart, is to know which elements contribute to making it a bad one. Here you can read about the elements that contribute to a bad pie chart:
1. Plain misuse of the chart
One of the reasons that the pie chart got its bad name is the plain misuse of the chart. In the graph below, you see that the data exceeds 100%. In no way the user is able to derive the right insights out of this visual.
Always make sure that your data adds up to 100%. This will definitely make your pie chart trustworthy, and viewers won't start asking questions about the integrity of your pie, and for a further extend, your whole dashboard.
2. Too many slices
When a pie chart has too many slices, it becomes very unreadable and the users simply cannot read all the values. A possible solution to this is to group certain categories in order to make it more clear. A good practice here is to keep your slices limited to maximum 6, which is already living on the edge, don't you think?
3. No labels
If a pie chart does not have any data labels, the user has no way (with exception to hovering if your tool allows) to see the data behind the slices. When data labels are applied, the values and slices become more clear. Often the percentage is displayed, but the user can choose to show the absolute values as well. Don't 'overdo' your pie chart with labels, values, percentages and what not. Determine what is the minimum valuable configuration, and the visual will stay clear to your viewer.
4. Random sorting
Sorting is very important for a pie chart as it eases the readability of the chart. Sorting the data clockwise (like you would read a clock) is seen to be a 'best practice' for pie charts. Some tools provide a rotation function to rotate pie charts, but please don't use this in any way, as it will only distort the visual.
Alternatives
The perfect alternative for a pie chart is a stacked bar chart which shows the data in percentage blocks. Each portion represents the percentage of the grand total. The user can read the data more easily and is able to compare the categories with more precision.
Conclusion
Pie charts. It seems that you will always have two groups of people that love them or hate them. In data visualization projects, a pie chart is often used by business and it will be difficult to exclude them as business often holds the money to complete these projects. Alternatives are available for pie charts like the stacked bar chart (100%), but sometimes it is also acceptable to show a pie chart for the sake of having a visual diversity of the dashboard.