Visualizing qualitative data—the non-numerical information collected through means of observations, interview, focus groups, surveys, and other methods—is undoubtedly a unique challenge. Results from interviews and focus groups can reveal many themes and sentiments, which makes summarizing the results difficult to do.
Yet, qualitative results allow us to build narratives and tell stories in ways that quantitative data can make difficult. Downloading a big data set, running regressions, and creating tables may provide more generalizable results, but readers don’t connect to them the way they connect to a story. Qualitative data can help tell those stories.
Here, I explain seven data visualizations that you can use to visualize your qualitative data. I’ve also listed a sample of tools that I have used to create each graph type (though there are certainly others).
In most cases, preparing the data for the visualization is going to be the hardest part—you are going to need to standardize the data, uncover the major themes and perspectives, and track down specific words or quotes. Several qualitative data tools can help you do this hard work, including NVivo, Dedoose, Qualtrics, Dovetail, and more. Many of these have their own basic data visualization tools built into the tool, but I don’t review those capabilities here.
Qualitative Data Preattentive Processing
Every data visualization, to at least some degree, relies on “preattentive processing”—or the automatic analysis of visual (or auditory) stimuli that occurs before conscious attention is directed towards the object. In other words, it’s the stuff you notice right away.
With quantitative data, we refer to attributes like the length of a bar in a bar chart or the position of a dot in a scatterplot. With qualitative data, we rely on a slightly different set of attributes to help differentiate the data. The image below is from Richard Brath’s book, Visualizing Text, and you can see how a single word (name) pops out of the group to grab and direct our attention.
In each of the seven qualitative data visualizations, these preattentive attributes can help you emphasize a point, pull out a key takeaway, or just create more visual interest.
Word Clouds
Tools: Flourish, Jason Davies, WordClouds.com
Perhaps the most popular and familiar way to visualize qualitative data, word clouds are really a way to display quantitative data: the number of times a word appears in a text. In a word cloud, the size of each word is adjusted according to its frequency in a passage. These graphs are probably best used to visualize overall patterns or to show where a single value is obvious and stands out. They are less appropriate in cases where finding specific values is important.
Word clouds are visually engaging, but they present two primary challenges. First, it is unclear what the specific frequency of each word is in the text. Second, the orientation of the words (along with their color and font) can affect our ability to identify frequent words.
Instead of one large word cloud, Marti Hearst and colleagues (2019) suggest breaking up the text into semantic categories (or “zones of meaning”) and creating multiple small word clouds instead. They find that participants in their study were able to identify categories more accurately and more quickly with this method. Of course, creating those categories requires more work than simply copying-and-pasting the full text into a word cloud tool.
Developed by Tommy Dang, Huyen Nguyen, and Pham Van Vung, the WordStream is a more recent innovation that’s similar to a word cloud, but shows changes over time and can be used to “communicate the global patterns of the text corpus across time.” Nguyen, Dang, and Kathleen Bowe have also created a free open-source tool that you can use to create your own WordStream.
Specific Words
Tools: Excel, Flourish, RAW Graphs
Another way to visualize qualitative text data is to combine individual words with a quantitative metric. I love this histogram-as-words from Matt Daniels at The Pudding that shows the number of unique words used by individual rap artists over time. Instead of a set of bars, the author plotted the artist’s name, using color to distinguish the decade of album release.
A related graph type is the beeswarm chart, which can be used to show individual points or words. I like this 2016 interactive piece from Nikhil Sonnad and Keith Collins at Quartz that “shows how well a word’s usage on Twitter corresponds to the level of Trump support in US counties.” It is a beeswarm chart, but the interactivity enables the user to see specific words.
Coloring Phrases
Tools: PowerPoint, Microsoft Word (track changes, duh!), R (ggpage)
Passages of text are composed of words, yes, but those words build phrases, which build sentences, and then paragraphs. Depending on the goal of your visual, highlighting specific portions of text may be useful to summarize your analysis. You can use visual design elements, such as color or boldface, to make important passages visible to your reader.
The New York Times, for example, highlighted the words and phrases in a week’s worth of public statements by then-President Trump to show his speech patterns. In the longer news story, the reader can dive into more detail to see the actual text from the entire transcript and the highlighted sections.
And the Los Angeles Times provided just a single image of the entire Mueller Report so we could see where and how many redacted sections were included.
Icons, Diagrams, and more
Tools: PowerPoint, Canva, Adobe Illustrator, NounProject
There are a lot of options here! Keep in mind that you might need to provide more text, but use icons or diagrams to make the entire piece more visual and more engaging. This short infographic from the Center on Budget and Policy Priorities, for example, is, at its core, just a bulleted list about how the Earned Income Tax Credit (EITC) and the Child Tax Credit (CTC) help families. Including the icons rather than abstract circles or no organizing feature at all, helps make it a bit more visual and engaging.
A word of caution: when using icons to represent people or communities, you want to be careful that you are representing real people who deserve your respect and empathy.
Heatmaps
Tools: Excel, Tableau, Flourish, Datawrapper, R
I’m generally a big fan of heatmaps where color and saturation of color are used to show values in your data. But heatmaps can also show categorical data, which might be perfect for your qualitative data. Of course, you are going to need to do some work to build the categories just as you would with other visualizations, but it may be worth it to give your reader an overarching view of the data.
Here is another example from the New York Times where they categorized elements of the GOP’s failed various efforts to repeal the Affordable Care Act. The heatmap shows what would happen under five separate Republican bills across a dozen different aspects of health care in the country.
Quotes
Tools: PowerPoint, Canva, Adobe Illustrator
Similar to the word clouds, showing direct quotes in a standard way to visualize qualitative data. There are three challenges with using quotes in the presentation of your data:
- Finding the quote or quotes that embody the spirit of the analysis. Does the quote accurately summarize the sentiment of the person or, more generally, all of the people represented in the data?
- Using a quote that is concise, easy to understand, and maybe even pithy. You don’t want to choose a quote that is 16 sentences long, but you also need to use something that accurately reflects the person’s perspective.
- Placing the quote in a way that looks good and is appealing to the reader. There are countless ways to lay out a quote on a page or slide or webpage, so I’ve presented a variety of examples below. If you’re struggling to find a way to design an image or slide for a quote, I would recommend trying the Canva design tool, the Designer feature in PowerPoint, or just a good ol’ Google search.
Combining multiple options
Finally, we can weave different visualizations together, which can be especially useful if you have quantitative data to pair with your qualitative data—maybe a quote about job loss with a graph on the unemployment rate. I really like this 2018 piece from the Texas Tribune in which the journalists explored where the Texas Democratic and Republic parties stood on six different issues. For each, they had icons at the top to separate that segment of the story, followed by a stacked bar chart showing results to a question from a University of Texas/Texas Tribune poll, a quote from each party’s platform document, and the journalists’ summary (“take”) on each issue.
And so on…
There are, of course, a wide range of qualitative data visualizations that are not included here: Venn diagrams, Histomaps, Fishbone diagrams, spectrum displays, and gauge charts are just a few. If you’re interested in learning more, I would recommend Chapter 10 of my Better Data Visualizations book, Richard Brath’s Visualizing Text, and Sheila Pontis’s Making Sense of Field Research.
Ultimately, visualizing qualitative data is the end of a long path of analyzing qualitative data. You may need to do a lot of work before ever getting to the visualization stage, such as creating transcripts of your interviews or focus groups, using data tools like NVivo or Dedoose to group and refine the text, or relying on tools or programming languages to help group, sort, or otherwise analyze the text. In the end, qualitative data often needs more work and care to find the underlying themes and patterns, but remember that such data can also enable you to relate and connect with your reader through more accessible stories.
Want to learn more about qualitative and quantitative data visualization? Check out my book, Better Data Visualizations, and my new virtual data visualization course with Skillwave Training, The Art and Science of Data Visualization.
