Making Healthy Choices about Data

Nik Eveleigh   by Nik Eveleigh, Principal BI Consultant


At a recent business analytics summit in Johannesburg, Synergy posed a simple question to the audience members via an app:

“Which do you prefer – a pie or a donut?”

As you would expect some people went for pies, others went for donuts and a small number sat on the fence. One respondent gave an answer along the lines of “it depends on the data” but I’m sure he or she wasn’t the only attendee to realise the question was about the representation of data rather than a shootout between fat and sugar.

Both the humble pie chart and its marginally more attractive donut-shaped sibling come in for a fair amount of (mostly justifiable) abuse. At face value however, either of them are decent representations of the question we posed, for a couple of key reasons:

  • Single set of data
  • Small number of categories
  • Comparing portions of the whole

Pie vs Donut chart

 

 

 

 

 

Neither of these visuals are going to set the world on fire but they do, at least, add up to a hundred and they do show that pies are a little more popular than donuts.

Let’s consider the why of the question and assume we wanted to know how prevalent the pie people were in the audience. Let’s also consider the where.

Illustrating how prevalent the pie people were in the audience. In answering the why and the where we can make a couple of design changes:

  • It’s at a Qlik conference so let’s give it Qlik colours
  • We want to know about pies so
    • Include a BAN (Big Ass Number) about pies
    • Group the donuts and the fence sitters together
…but when all is said and done do we really need any fat or sugar to answer the question? The percentage of people that prefer pies to donuts

Whether we’re talking about food or data, the occasional pie or donut isn’t the end of the world – but it’s always good to be aware of a few healthier alternatives.

Healthier alternatives

Continuing our food theme, let’s ask a question to the donut brigade – “What’s your favourite donut?” – and display the top ten answers as a pie. Still with me? Great.

What’s your favourite donut?

 

 

 

 

 

 

Here’s the same data on a bar graph displayed in descending order based on percentage of the vote.

Using a bar chart to display the data

 

 

 

 

 

Even with the same hideous colours, the data is a lot clearer in the bar chart and it’s much simpler for the consumer (pun very much intended) to compare categories.

Let’s take a similar example but this time with an additional grouping. We’ll ask Team Pie – “What’s your favourite pie filling?” – then categorise the answers and display the top ten. As before we’ll go with a pie chart versus a bar chart.

Pie charts and categorisation

 

 

 

 

 

 

 

Favorite type of pie filling Bar chart

OK so we’ve moved from a psychedelic colour scheme to variations on a theme* – blues for savoury and oranges for sweet. Once again, the bar chart is far simpler to make sense of in the context of the question.

*before the design police arrive let me please state for the record that I’m not promoting unicorn puke or fairly meaningless colour gradations on a bar chart, I’m just trying to make comparisons on a fairly level playing field…

All the cases so far have examined categories of data compared to each other. Let’s bring time into the equation and assume we polled the excitement levels of 100 people in relation to Synergy before and after this blog post.

Comparative pie chart anyone?

 

comparative pie chart

 

 

 

 

 

A stacked bar chart perhaps…? Comparative bar chart

 

 

 

 

I’d say the smart money should be on a slope chart using colour and annotation to convey the required message about the data…

Slope chart

 

 

 

 

 

 

 

 

Miles Kingston famously wrote:

Knowledge is knowing that a tomato is a fruit
Wisdom is not putting it in a fruit salad

At Synergy, we apply our knowledge and wisdom to help people make better decisions.

If you are looking for more intuitive ways to represent your data, or to better monitor and steer your business profitability using your data, we would love to hear from you.

If you liked this Data insights article, and would like to subscribe to this data insights series, you can subscribe here:
(You can unsubscribe at any time).