Last year, I was asked to write two “Rip Outs” for the Journal of Graduate Medical Education. What is a “Rip Out”, you ask? Rip Outs are purposely short articles designed to provide readers with information to facilitate their work as educators in graduate medical education. They are, in the words of the journal, “intended to be ‘ripped out’ (or downloaded) from the Journal.”
Each Rip Out follows a specific four-part structure:
- The Challenge. This is the thesis statement, summarized in one paragraph.
- What Is Known. This is the current state of the science or practice, again, just one paragraph.
- How You Can Start TODAY. Here, the author lays out the short-term knowledge, practice, or strategies the reader can take immediately to integrate the thesis into their work.
- What You Can Do LONG TERM. Here’s where the author suggests ways to extend the lessons from the article over the long run—tools to use, classes to take, books to read.
The Rip Out just published, co-authored with Meihsi Chiang at Washington University in St. Louis and Bo Kim at Harvard Medical, focuses on qualitative data visualization. While most data visualization tools—and most data visualization advice—are built for numbers, a huge share of the information that drives decisions shows up interviews, focus groups, narrative assessments, and open-ended survey responses. That kind of qualitative data can be more difficult to quickly summarize and visualize, but deserves the same design attention we give to bar charts and line graphs.
In our new article, “Visualizing Qualitative Data: Challenges, Strategies, and Opportunities,” we define four categories of qualitative data—textual, audio, video, and image—and how each resists the standard dataviz toolkit of bar, line, and pie charts.
To effectively communicate qualitative data, we encourage readers to understand their audience—what they need to know and what platforms or visuals they are most likely to use. We urge them to use any tool that will help them build their visuals, from Flourish to PowerPoint to Canva. And we try to help readers understand that data visualization is a mixture of science and art, and that there is usually not a “right” and “wrong” way to visualize their data.
Ways to Visualize Qualitative Data Graphic
The main graphic in the article—which takes up about a half-page of the two-page article, so that was a challenge—covers eight approaches to visualizing qualitative data—including word clouds, quotes, colored text, timelines, flow diagrams, icons and Harvey balls, heatmaps, and gauge/bullet/slider charts—each with a quick note on when it works and where it falls short.
We didn’t want to stop at just those few qualitative visualizations, so we also created a longer graphic with more examples and more details that you can see below and download (for free) as a PNG or PDF for your use and reference. You can also access the full article (again, for free) at the JGME website.
If you work with narrative, interview, or evaluation data and have been reaching for a bar chart out of habit, I hope this gives you a few new places to start.
