Chart choosing is part art and part science. During workshops, I cover my chart choosing thought process in more detail. For now, let me save you hours of time with this not-so-secret secret: Start with a bar chart.
Then, fine-tine your bar chart. Try your numbers as a regular bar chart and as a stacked bar chart. There are subtle differences between the two options and it’s worth discussing the pros and cons of each approach with your coworkers before publishing your work.
Option A: A regular bar chart
A bar chart should be your go-to chart. No, not a donut, a radar, or a spaghetti line graph. Those are tempting options, I know! Stick with the basics. Don’t distract your viewers with elaborate graphs. Let them look past the graph and think about what the numbers actually mean.
Option B: A stacked bar chart
Light gray shading reminds your viewers that the blue section is only part of the story. No, I didn’t label the gray section. Yes, that was on purpose. Keep the focus on what matters–the blue.
Five years ago, when I was an in-house data person for a large nonprofit in DC, the regular-versus-stacked-bar-chart conversation went like this.
Me: “Hello there, High Up Manager. I’ve got drafts of that important slideshow for the Board to share with you.”
High Up Manager: “Why thank you for being so prepared. I always appreciate a number-cruncher who takes time to look at the numbers in a lot of different ways.”
“My pleasure. Here’s the first visualization: A bar chart showing the number of youth in our programs that graduated from high school on time. Tell me what you think.”
“Wow! We’re doing great! Especially the Category A youth! I wasn’t sure if we were on track! We won’t have to change our programmatic approach at all because everything is going swimmingly!”
“Okay. Now here’s the second visualization: Another bar chart, but this time, the number of youth that didn’t graduate on time is shaded in light gray. Your thoughts?”
“Hmm. Does that gray section show the students who didn’t graduate on time? That’s a pretty big proportion of the graph! I had a feeling that there were changes to make. This is a good reminder that there’s still room for improvement. This is the version we need to show the Board at next month’s meeting. But first, let me run down the hall and share the numbers with our program staff so we can start acting on the data.”
Verbatim? Of course not. The gist of what you should be discussing with your teammates? Absolutely. There are a dozen correct ways to visualize every dataset, and the regular vs. stacked conversation should be part of your drafting process.
Antoinette
5 Comments
Always so useful, thank you Ann! I’d love to get your take on a burning chart labeling question.
I know that bar graphs should remain unbroken and stretch to the 0 axis in order to be interpreting accurately by our brains. But I have seen data labels placed inside the end of the bar. Sometimes I like this look for aesthetics / space saving purposes. But I can’t find a definitive answer on whether that is an optimal practice. Your thoughts?
Thank you for all you do!
Great question. I was just talking about this during a workshop today. I’ll blog about it one of these days!
Hi Ann I’ve been watching your videos on YouTube and find them very interesting. I have some questions about analysing surveys and I would like to make direct contact with you. I would very much appreciate your response.
I’m looking forward hearing from you.
Thanks
Regards
[…] I built stacked bar charts instead of regular bar charts because I wanted to remind viewers that while some people checked the survey boxes (dark blue) others did not (light blue). I only labeled the dark blue bars because that’s the segment that really matters. […]
[…] been showing you one tip at a time, like deciding how far you’ll stretch your scale, whether you’ll use a regular or stacked chart, and whether your chart will be oriented vertically or horizontally. Today, let’s check out […]