Pie Charts
Pie Charts
Embedded Files
A pie chart is a common univariate plot type that is used to depict relative frequencies for levels of a categorical variable. Frequencies in a pie chart are depicted as wedges drawn on a circle: the larger the angle or area, the more common the categorical value taken. Use a Pie chart only when the number of categories is less, and you'd like to see the proportion of each category on a chart.
A pie chart is a common univariate plot type that is used to depict relative frequencies for levels of a categorical variable. Frequencies in a pie chart are depicted as wedges drawn on a circle: the larger the angle or area, the more common the categorical value taken. Use a Pie chart only when the number of categories is less, and you'd like to see the proportion of each category on a chart.
Pie chart (left) and bar chart (right) displaying the same categorical counts, 'generation_id' in the pokemon dataset.
Pie chart (left) and bar chart (right) displaying the same categorical counts, 'generation_id' in the pokemon dataset.
Unfortunately, pie charts are a fairly limited plot type in the range of scenarios where they can be used, and it is easy for chart makers to try and spice up pie charts in a way that makes them more difficult to read.
Unfortunately, pie charts are a fairly limited plot type in the range of scenarios where they can be used, and it is easy for chart makers to try and spice up pie charts in a way that makes them more difficult to read.
Guidelines to Use a Pie Chart
Guidelines to Use a Pie Chart
If you want to use a pie chart, try to follow certain guidelines:
If you want to use a pie chart, try to follow certain guidelines:
- Make sure that your interest is in relative frequencies. Areas should represent parts of a whole, rather than measurements on a second variable (unless that second variable can logically be summed up into some whole).
- Limit the number of slices plotted. A pie chart works best with two or three slices, though it's also possible to plot with four or five slices as long as the wedge sizes can be distinguished. If you have a lot of categories, or categories that have small proportional representation, consider grouping them together so that fewer wedges are plotted, or use an 'Other' category to handle them.
- Plot the data systematically. One typical method of plotting a pie chart is to start from the top of the circle, then plot each categorical level clockwise from most frequent to least frequent. If you have three categories and are interested in the comparison of two of them, a common plotting method is to place the two categories of interest on either side of the 12 o'clock direction, with the third category filling in the remaining space at the bottom.
If these guidelines cannot be met, then you should probably make use of a bar chart instead. A bar chart is a safer choice in general. The bar heights are more precisely interpreted than areas or angles, and a bar chart can be displayed more compactly than a pie chart. There's also more flexibility with a bar chart for plotting variables with a lot of levels, like plotting the bars horizontally.
If these guidelines cannot be met, then you should probably make use of a bar chart instead. A bar chart is a safer choice in general. The bar heights are more precisely interpreted than areas or angles, and a bar chart can be displayed more compactly than a pie chart. There's also more flexibility with a bar chart for plotting variables with a lot of levels, like plotting the bars horizontally.
Answers
Recommended Read
Recommended Read
Further Reading
Further Reading
- Eager Eyes: An Illustrated Tour of the Pie Chart Study Results - how accurately do people perceive different formulations of the pie chart?
Supporting Materials
Supporting Materials
Google Sites
Report abuse