Pie Charts in Statistics Explained

In the previous segment, you learnt bar graphs or bar charts as a form of data visualisation. Another handy data visualisation tool is the pie chart. 

Pie charts are a visually appealing and effective way to understand how specific categories in your data contribute to the whole. In layman’s terms, they are similar to visualising how much a pizza slice contributes to the whole pizza. 

In the forthcoming video, you will see the pie chart in action. 

We recommend you to open Excel and follow the steps in the video to understand data visualisation better.

Pie charts have limited uses in the representation of data. Although they can help you visualise the composition of specific groups in a data, they do not reveal the absolute value of each group. Hence, they need to be supported by adequate data labels to depict such information. 

Further, they are best used when the categories are fewer, as that would result in substantially larger and easily interpretable pieces of the pie. Consider how easy it is to split a pie (or a pizza) into 4 or 6, or 8 pieces, rather than 20 pieces or more.

There is one workaround in this case, depending upon how the data is distributed. Suppose you have 20 categories in your data, of which 3 are formidable chunks, which constitute substantial pieces of a pie, whereas the other 17 are quite small chunks, which constitute pie pieces of almost hairline width. You could club all the 17 pieces as ‘others’ and easily represent it as a fourth piece. But what would you do if each of these pieces is as important as the others?

Say, for example, you are measuring the average salary of the employees in an organisation. For 2,000 employees, the average salary comes out to be $850, with the 10 CXOs earning an average of $50,000. Obviously, the CXOs are too small to be represented as one category; due to this, you might be inclined to club them with some other categories and name the entire new category ‘Other’. 

Here is a pie chart to help you visualise this situation. 

Now, suppose the CXOs receive a 12% pay cut due to the poor performance of the company. This would reduce their overall contribution to the computation of the average by $60,000, causing a reduction of $30 in the average employee salary.

And in this case, the pie chart would be rendered insufficient to explain why there are changes in the average while the composition of the employees has not changed. The insufficiency of the pie chart in depicting how the data is distributed across multiple categories and how the average varies according to changes in the minor categories of the data leads us to find a different mode of visualising data: Histograms. You will learn about histograms in the next segment.