If you took chemistry in high school, you may remember doing a litmus test at some point. It’s when you dip a small strip of litmus paper into a solution, and the resulting color reveals whether it is an acid or base. In other words, a single factor determines the finding. There is a similar test that you can perform on your data visualization to determine whether they are good or bad. And here we introduce you to the data visualization test:
You can easily translate visual data into simple, semantic insights with this data visualization test
Let’s look at an example.
Here we are looking a count of oil pipeline accidents, organized by Cause Category and Cause Subcategory. It is a stacked column chart, but showing each count as a percentage of the overall total. So, can you easily translate any of this into plain English? No. This chart fails the test. Now look at this one:
Here are some simple insights that you can gain at a glance:
- Accidents are trending up
- 2011 had the fewest accidents
- 2016 had the most accidents
This one clearly passes the test. It’s fairly basic, but it is simple and tells a clear story. Furthermore, it is actually quite easy to take a non-sensible visualization and turn it into something clear. Going back to our first visualization, we can remove the percentage distribution:
And then remove the legend values:
Now can we easily translate this visual data into simple, semantic insights? Absolutely. The next time you’re reviewing some charts and graphs for your business, apply this simple test to see how your results can be improved.