A few weeks ago, I needed a classification of chart types for my book, and reinventing the wheel was the last thing I wanted to do. I started with Andrew’s classification and the Juice Analytics version. It’s a good starting point, but I couldn’t fit it into my work, so I decided to design my own classification (it’s always nice to have new standards), inspired by this one but also by others, like S. Few’s Chart Tamer. And no, I was not going to design a new periodic table of data visualization methods.
Here is the idea: a chart can (a) help you compare data points faster but keeps each data point as the basic information unit or (b) help you generalize the data and find patterns, making the data points less relevant. These roles should be as mutually exclusive as possible (but in real world that’s harder than expected).
There are six types of questions, three of them involving data comparison and the other three data reduction. The questions are:
- Comparison: comparing and sorting data points;
- Composition: part-to-whole comparisons;
- Distribution: comparison of data points along an axis;
- Relationship: relationship patterns between two or more variables;
- Evolution: time patterns;
- Profiling: pattern comparison.
- The first row lists the most commonly used chart in the category (it’s a fact, not something I agree with);
- Only charts that can be made in Excel are included;
- Depending on the question and the task, a chart type can be found in more than one category;
- You may find it bizarre to list a dot as a chart type, but one/off dots are critical in some tasks (monitoring) so I though they deserved to be included;
- There are no neutral and objective taxonomies.
Please share your comments and suggestions below.
(click to enlarge)