When it comes to choosing the right chart for your data presentation, the sheer variety of options in modern analytics software can feel overwhelming. So how do you decide which one will effectively convey your message?
With the technology available today there’s a risk of being drawn into a style-over-substance approach to visualisations. And with so many ‘pretty’ infographics available, there’s a temptation to switch things up and avoid a more simple approach.
Some data visualisation “gurus” encourage the use of infographics and stylised charts, often citing examples from the media. But it’s important to remember that media graphics are designed primarily to grab attention rather than to communicate complex information in an unbiased way.
But why choose basics when other options might seem more visually appealing?
This question led myself and a colleague to research our long held theory; simple works best!
Researching the theory
In developing my Data Visualisation workshops, I spent time researching key publications from well-known experts in this field, and there are quite a few! People like Stephen Few, Colin Ware, and Cole Nussbaumer Knaflic have all contributed significantly to the landscape of visualisation knowledge, and their books are both enlightening and informative. However, after reading these works, I was left wondering, “The theories make sense, but do they hold up practically?”
In his book Information Visualisation: Perception for Design, Colin Ware discusses preattentive attributes and how visualisations that incorporate these attributes can appeal to our iconic memory, helping us process information without much conscious effort, thereby allowing us to understand it more quickly.
Testing the theory
My curiosity to put these theories into practice sparked a conversation with my then-colleague Andy Mills, an experienced analytics leader who is deeply committed to the professional growth of analysts. Together, we launched an analytics academy within the Demarq group to support ongoing development in the field. While discussing content for an upcoming Data Visualisation event, we decided to take it a step further by partnering with Lancaster University in the UK to rigorously test these ideas.
As part of this project, Master’s students at Lancaster University developed an experiment where participants were shown various chart types displaying the same data, then asked questions based on what they were viewing.
Unbeknownst to the participants, they were seeing the same data represented in different chart formats throughout the test. The test was conducted on laptops, with software recording how quickly and accurately participants answered questions. Speed and accuracy served as indicators of each chart’s effectiveness. For example, a participant might see a chart showing the market share of different products for a fictional company, then be asked to identify the product with the highest or second-lowest market share.
The results
By experimenting with different chart and data combinations, we set out to determine which visualisations most effectively conveyed information with speed and accuracy. The results were decisive: in nearly every scenario, line and bar charts emerged as the top performers, reliably delivering information both quickly and accurately.
This completely aligned with Ware’s (and others’) theories. Ware suggests that when presenting numerical data, visualisations that use preattentive attributes like spatial positioning (in line charts) and length (in bar charts) work better than other attributes, such as contrast, motion, or size.
These findings reinforced a belief I've held for years: line and bar charts should be the go-to choices for visualisers.
Stick to boring - it works!
So, don’t get seduced by shiny, pretty charts! Stick to the familiar and scientifically proven approach of using lines and bars - they’re rarely bettered.
While I acknowledge that this isn’t the only way to approach visualisation and may seem ‘boring’, I can tell you it’s served me well over my 35-year career!
Steve Hulmes - Analysts Coach, Sophic
If you would like further support to create charts that hit the mark and deliver your analysis clearly to your audience you can access some effective strategies now by subscribing to my Newsletter on LinkedIn ‘The Analyst: Insightful Bytes’ and reading edition 6. Just follow the link:
To find out more about our data visualisation workshop and to view Sophic's full programme click here.
However, if you’d like to discuss a more bespoke solution to your analysts’ development needs, let's have a no obligation chat to see whether Sophic can help. Just email Steve at steve@sophic.co.uk
Comments