Dashboards That Pop: Pie Charts
Shows: Parts-to-a-Whole and Proportions
Pie charts are frequently misused, so some care is needed to ensure that you are creating pie charts that are useful. Pie charts should be used when comparing each slice’s contribution to the whole as opposed to comparing each slice to each other.
If we look at this pie chart, we can easily see that the red slice contributes very little to the whole, meaning that overall, our dataset contains very little red category data. Trying to compare the red slice to the orange slice is another situation entirely. Is one slice bigger than the other, how much bigger is it? Are they the same size? It is hard to tell.
Accurately judging the size of pie slices is difficult to do and is even harder when trying to do so across a series of pie charts.
So now that we know that pie charts should only be used for comparing a given category (one slice) within the total of a single pie chart, let’s explore configuration best practices.
Let’s start with a default pie chart and walk through how to apply best practices.
Best Practices
Limit the number of categories
Shoot for five or less categories. The more categories you have, the harder it is to see the smaller slices. It is also difficult to select enough colors that are easily distinguishable from one another.
Our default pie chart has too many categories. We can resolve this by focusing on the larger categories. Do not try to reduce the number of categories displayed by filtering your data, as this will distort your data. Remember the entire circle represents 100% of your dataset. If you filter, you are removing data from your 100% which will skew the results. Instead, we want to group the smaller categories together into an ‘other’ slice.
The default and grouped pie charts show Riots/Protests to be 28.48%, while the filtered pie chart instead displays it as 32.08%.
At the bottom of the Slices tab, you will find a Groups (%) box. Experiment with different values until you find the value that works for your dataset. Any individual categories that are below that grouping value will be grouped together. In the example below, the grouping value is 6.5%. All categories that make up less than 6.5% of the data will be grouped together.
Sort
To make it easier for users to understand which slices are larger than others we will sort our slices by size. This way, the order of the slices helps users determine which is larger than the other.
At the bottom of the Data tab use the Add field dropdown to sort by Statistic. Descending positions the largest slice at the beginning, while ascending starts with the smallest slice.
Color
Each slice should be a distinct color that is distinguishable from any other colors in the pie chart. Also ensure that you are choosing color blind friendly colors.
There are a variety of browser extensions that can simulate color blindness to help you assess your dashboards. To learn more about how to effectively use color in your dashboards, check out Dashboards That Pop: Color.
Sizing
Make sure you are allocating enough room for your pie chart. If your users can’t easily see and read it, why bother? If you are counting on them to expand it, they are most likely not going to do that.
Aesthetics
The following tips and configuration options really come down to designer preferences.
Legend vs Labels
While I am personally on Team Legend, the choice is yours as to whether or not you want to use labels or legends. You can also choose for either of them if you want to display the value or the percentage. The choice may also be dictated by the space allotted to the pie chart. Labels can be really difficult to use if your pie chart element is too small.
While the choice is yours, you do have to make a choice. Having neither force users to have to rely on the hover (if enabled) function. This isn’t a great user experience. Having both is just a confusing mess.
The label, legend, and hover options are located towards the bottom of the chart tab.
Donut Chart
A donut chart is a pie chart with a hole in the middle. Some say that it is easier to judge the slices of a donut chart, as you are no longer trying to judge the area but are comparing the length of the arc, while others say there is no difference between the two. So, I will leave it as a design preference for now. That being said, I do prefer the donut chart as it looks a bit more modern to me.
To donut-fy (that is totally a word) your pie chart, experiment with the Inner radius value in the Data tab.
Here is the evolution of our pie chart.
Dashboards That Pop: Inspiration
Dashboards That Pop: Survey123