Data Visualization & Chart preparation
Visualization is a powerful tool that can do wonders when used appropriately.
Few days back, I attended a presentation, in which the presenter threw a couple of charts at the audience which were very similar to the one below:
It was more like a hit-and-run presentation. The moment I saw these charts, I decided to do some research and write an article on how a chart should NOT be.
So here are some of my insights that can leverage the quality of the charts and presentations :
- Make it simple
- Avoid Pie and 3D charts
- Make charts speak for themselves
- Use blank spaces strategically
- Avoid clutter
- Drive your audience’s attention
- Keep your context clear cut
In addition to that, I have prepared a cheatsheet to select appropriate chart type. It can be downloaded at the end of the article.
Make it simple
Simplicity is the ultimate sophistication. Minimalism is the new trend.
Audience should never feel the difficulty to read the chart. A good chart should induce discussion on the context rather than the chart itself. Follow the concept of 5-seconds. If your audience take more than 5 seconds to understand the visual, the chart fails.
Avoid Pie and 3D charts
Worst thing a data visualizer can do is to allow pie charts and 3D graphs into his presentations. Human brain cannot interpret the sizes of the slices of a pie chart, leading to misinterpretation of the data. 3D charts will add clutter (more on clutter in latter topic) to the visual, making readers hard to understand the data. Chart making softwares (like MS Excel) still have 3D options, tempting us to opt for it. I am not asking to completely ignore the Pie and 3D charts. Pie charts can be used only if the number of variables to be displayed is not more than three and difference between them is very high.
Look at the below pie chart. We cannot conclude if TCS or Capgemini have higher employee frustration rate without markings as our brains are not designed to perceive the sizes of pie-slices. (context for these charts are just for fun)
Take a look at the below 3D chart:
Readers have to project the end of every bars to the plane behind the bars and track to y-axis for readings. Imagine how hard it will be.
My worst nightmare will be blending 3D with Pie charts. Like this :
In the above visual, IBM and CTS almost looks same size. A completely wrong data would be passed to audience if there is no markings. This is the reason why these kind of charts should be completely avoided.
Make charts speak for themselves
Imagine your chart as a real person. How do you feel about a person who is not speaking for himself and need someone else to express his feelings? Same applies to the visuals too. An image is worth 1000 words only if the main subject of that image is instantaneously understandable. Look at the below image that won Pulitzer prize :
No captions needed... No explanations needed… The image speaks for itself.
Although the same perfection cannot be replicated but anyone can make a visualization that can speak for itself without someone standing there explaining it.
Wherever geographies are involved, better use the map of that particular country or region and plot the data.
Use blank spaces strategically
There is common thought that if there are lots of white spaces on the slide, it means the content is less. Blank spaces on the chart don’t have anything to do with the content richness of the chart. Its not important how much space you filled in the chart but how much constructive discussions the chart generate in the audience is important.
If you are about to display only a single value or a percentage, then no need to display it on a chart. Just show the numbers directly like this :
Avoid clutter
Anything on the chart that takes up space but don’t add any value to the chart is called clutter. Clutter increases the cognitive load on the audience. They will take some time to understand the chart and decide where to focus. As a designer, this task can be done by us and can deliver clutter-free chart to the audience, reducing their cognitive load and easing up their lives.
Following images shows the same chart before and after clutter removal process :
One more important thing to note is that the orientation of x-axis text has been changed. In a study by Wigdor & Balakrishnan in 2005, it has been found that any text rotated to 45 degrees in any direction will slow down the reading speed by 52%. If it is rotated by 90 degrees, reading speed will be reduced by 205%. Read their paper here
Drive your audience’s focus
When you are showing something to your audience, then its your duty to drive your audience’s eyes and control where they should see on the chart. You are the one responsible if they take away their eyes from the chart and peep into their Blackberry’s.
Charts should be appropriately designed that if someone look at it, their eyes should automatically be drawn to the context of the chart.
For example, look at the below heat map of India where climatic conditions are plotted :
Viewers attention will be immediately grabbed by the areas marked on the legend. No need for someone to drive the audience’s attention.
Keep your context clear cut
I saved the most important ingredient of chart making recipe for the last. We have seen hundreds if not thousands of movies without a valid story line but managed to make huge sum in box office. Unfortunately business presentations are not movies. Here the visualizations must be accompanied by a strong story line.
Before sitting down to prepare a presentation, ask yourself the following questions to get a clear idea of what your context is :
1. Is my presentation exploratory or explanatory ?
2. Who are the audience ?
3. What should be conveyed ?
4. How it can be delivered ?
Keeping these things in mind, one can prepare a decent chart and can tell a story with visualizations. Check this link to learn how visualization can be used for exploratory data analysis.
Recommended readings
1. Story telling with Data by Cole Nussbaumer Knaflic. An interesting and informative book thoughtfully written by Cole. She worked with Google and developed her own Data Visualization course there. Investing in this book (around 2000 rupees) would be a best starting point for anyone who wants to know more about Data Viz. Cole also pens a blog storytellingwithdata.com
2. Good Charts by Berinato
Check out the below infograph for effective preparation of a visualization :
Download the PDF version here
Read my other articles :
