Visualizing multi-select survey questions (where the respondents can provide multiple answers) is challenging to summarize since pie charts and 100% bar charts do not work well since the parts can add up to be more than 100%. In the book Picturing the Uncertain World: How to Understand,
Communicate, and Control Uncertainty through Graphical Display, the author gives the example of a survey of sermon topics, and uses a variant of Nightingale– rose (also called a polar area chart or “coxcomb”) to display the survey results:
The author, Howard Wainer, comments that this chart has a flaw — while the length of the filaments is proportional to the size, there’s an implication that the reader is interpreting it as an area chart (as would be in the Nightingale rose which would use solid shapes, rather than filaments). Here each data point uses 9 filaments of varying lengths. Here’s an alternative way to visualize the information using a simple bar chart:
Finally, a matrix view. Since I didn’t have the original data, I added 10 groups of respondents and also rounded to the nearest 10%. The rows contan the same data points as the original, but the columns show, for example, a group A which only preaches on two of the topics, while group B preaches on seven /10 topics. These lend themselves to additional grouping and insight. Perhaps goups B and C emphasize more social issues in their sermons, while group A focuses on other types of topics. This matrix view is more information dense (presents more data) and can yield more insights depending on the survey.