The Big Book of Dashboards

Chapter 34 The Allure of Pies and Donuts

It's easy for us as teachers, writers, or speakers to discuss data visualization best practices, but we also realize there will be times when you will not have complete control over design decisions. There may be situations where you cannot avoid using pie and donut charts. A client request or a demand from the Director of All Things Circular may force decisions to be made about a visualization that are not best practice. It is our hope that this chapter will help you in these situations.

Background

As discussed in Chapter 1, the two best encoding methods for making precise quantitative comparisons are (1) using length or height from a common baseline for the comparison (e.g., a bar chart) and (2) using position to make the comparison (e.g., a dot plot).

When trying to show precise quantitative comparisons, using angles, arcs, area, or size of circles is not as good as using length or position to encode data. For this reason, pie charts, donut charts, and bubble charts typically are not good choices for visualizing data. There are the occasional exceptions, but be very cautious in your use of these charts. Some examples follow.

Pie charts may be useful on a map to show a part- to-whole relationship within a geographic region. (See Figure 34.1.) This is because there is no easy way to present multiple bar charts on a map where there is no common baseline for making comparisons.

Figure 34.1 Pie charts on a map showing closed (blue) and ...