Chapter 62. The Case for One-Dimensional Unit Charts

This year I have found myself gravitating toward stacked bar charts as a way to show comparisons. The chart type is featured prominently in both my NBA Records by Player and 50 Years of AFC vs. NFC Matchups visualizations. I even used a stacked bar chart as the primary visualization to share how much progress I was making while writing Practical Tableau.

Something is different about these stacked bars though. I’m generally not a fan of stacked bars because if the stacks are different sizes, and if the stacks do not sit on the baseline, it can be challenging to compare and contrast the different pieces of the bars.

In all three cases I just mentioned, the stacks have equal units. This removes the challenges I have with traditional stacked bars, so I do not think the name gives these charts justice. So that got me wondering if these should be called unit charts.

If you’re not familiar with unit charts, here is a well-done example from one of my data viz heroes, Andy Cotgreave. My issue with the multiple columns and multiple rows approach to most unit charts—especially when they are showing a part to whole relationship—is that these are pie charts in disguise. Very sneaky. Andy always does an amazing job of eliciting dialogue around data visualization in the community so I assume he had a reason behind his chart type selection. Even if he didn’t, I’m not mad at him; we all have guilty pleasures. I once made a donut chart… and I ...

Get Practical Tableau now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.