Why you need to stop using Pie Charts when presenting visual analytics

Pie charts are often a popular choice among people to help convey visual analytics in a simplistic way.  We understand why, as on the surface, its a simple graph.  The truth is although people are craving simplicity in their visual analytics, pie charts aren’t the most effective way to present information.  In this article, we outline why you need to stop using pie charts and start embracing other ways to present your information in a clearer way.

Pie charts make processing information slower

It may be assumed that the aim of a pie chart and the simple design makes a large amount of data easier to process.  However, this is proven to be in fact the opposite.

Take the 2016 US Election results as an example;

Why you need to stop using pir charts

Donald Trump won the 2016 Presidential election, however the total number of popular votes shows that Hillary Clinton was the favourite.
Source: http://edition.cnn.com/2016/12/21/politics/donald-trump-hillary-clinton-popular-vote-final-count/

Click here for an interactive version of the US Popular Vote graph.

While Trump won the election, he lost the popular vote by nearly three million votes!

When you look at this pie chart, you see straight away that the race was close and many of you may have been able to pick out that the blue slice was bigger than the red.  Fortunately, the labels give us the exact numbers.

We use a pie chart when we want to show the comparison of several things, but if there are a lot of items to compare, space becomes an issue.  For example, this comparison of the population of each country is limited:

Why you need to stop using pie charts

* Rest of World represents 224 countries with populations ranging from 128,632,004 down to just 801 (median = 4,827,299)

Source: http://www.worldometers.info/world-population/population-by-country/

Click here for an interactive version of the World Population graphs.

Pie charts can be inaccessible for individuals with a vision impairment

Interpreting pie charts requires the user to be able to determine between colours and to have a good standard of vision.  However, according to self-reported data from the ABS National Health Survey, nearly 12 million Australians (53.7% of the population) have one or more long-term eye conditions and vision impairments (e.g. partial blindness, reading glasses, colour-blindness).

Use this checker to see if your colour scheme is accessible

Pie charts as visual analytics can be misleading for the brain

When you want to show a comparison, you really need your audience to see that in an instant.

But because pie charts are circular, the human brain struggles to accurately compare the size of the segments when they are at different angles and when there is no scale.  This means that most pie charts are misleading, acting like an optical illusion and distorting what’s being processed in our heads.  In short, our brains need to work hard to figure out pie charts and sometimes, they simply just guess the size of each slice.

When pie charts are in 3D, it makes it more misleading to process and some companies such as Apple have taken advantage of misleading their audience with a pie chart.

The principle of Gestalt perception outlines how the brain perceives and processes things we see.  The theory called Gestalt psychology is a set of laws that explain how we make conclusions from visual objects; why we take short cuts.

If you want to use a graph to show comparisons, you must follow the eight laws that makeup Gestalt’s Principles of Grouping.

Or you could just use a bar graph.

A better way to visual analytics: Using a bar chart instead of a pie chart

Let’s take the data from the US Election popular vote pie chart and put it into a column chart.  The numbers are still the same but this time, the comparison is plain to see.

Click me - US Popular Vote 2016 Bar Chart

And when there are many categories, a bar chart is ideal

Click me - World Population 216 Column Chart

Unlike a pie chart, you can clearly see every country, and you can even pick out the slight difference between China and India.

The bar chart is unambiguous.  It is very easy to compare the size of the bars; each country represented clearly, and no detail is lost.  You can even scroll down the bar graph and see all 234 countries.  Try and squeeze all 234 countries into a pie chart and I’m sure you’ll be breaking all of Gestalt’s laws and more.

So, the next time you think about creating a pie chart for data visualisation, think again.