A big part of our work here at Circyl is to help our clients source, manipulate, analyse and share valuable information across their organisation to deliver a real-time view of performance and productivity.

The ability to visualise these insights allows business users take action and drive better outcomes; all of which relies on presenting data in a way that can be easily interpreted. When data is presented using a suitable visualisation type it allows decision makers to quickly grasp difficult concepts or identify new patterns.

Unfortunately, the desire to present data in a graphical format can be where the trouble begins. In the world of modern data visualisation tools there is still a place for the classic pie chart. Now don’t get me wrong, for the quick comparison of two data points, such as the split between those in the office that prefer tea to coffee shown below, the pie chart does have its place, but their consistent overuse is often misguided.

Familiarity breeds contempt

The reason pie charts are so popular is that they come as standard in many popular office applications like Microsoft Word and I‘m confident everyone has at one time or another included one as part of a report or presentation. They’re easy to create and can brighten up dull pages or board reports with table after table of numbers. They have even evolved into several varieties to improve their appeal (3D, doughnut or exploded) but the basic fault remains the same; the more data you want to include, the harder they are for the brain to process quickly.

If I expand on my drink preferences chart and merely ask people in the office what is their preferred choice, rather than just tea or coffee, the result might look something like this:

We have now added more data points which may well offer an insight into the lifestyle choices of the individuals within our office, but it’s not as easy to compare the results by merely looking at the segments.

Which is more popular, water or a fizzy drink, squash or alcohol? Did any tea drinkers switch to another preference? It’s not easy to compare now that the segments are relatively similar sizes so we add the data labels to the segments to make it more obvious how many would prefer alcohol to help them through the day:

Now it becomes clear that alcohol is the least popular choice, but it is the labels that have delivered the useful information so why not just present the numbers in a table?

If you want a visual depiction of the results then a column chart is far easier to interpret, particularly as we add even more data. Switching our data to a column chart and ordering in terms of preference makes it easier to understand and compare against each other at a glance even without data labels on the individual columns.

Hopefully this brief example has highlighted the importance of choosing the right visualisation type for the dataset being presented and in the second part on this topic I’ll explore alternative options to the pie in more detail and the impact it can have on your business insights.