12 Comments

    Leave a Reply

    Your email address will not be published. Required fields are marked *

    How to Visualize Qualitative Data

    Sep 25th, 2014 / Data Visualization / , , , , , , , ,
    Visualize qualitative data from interviews, focus groups, surveys, and more by using word clouds, photographs, icons, diagrams, and timelines.

    Are you looking for ways to display your qualitative data? The vast majority of data visualization resources focus on quantitative data. In this article, let’s look at some of your options for qualitative data visualization, like word clouds, photographs, icons, diagrams, and timelines.

    I used to conduct qualitative-heavy research projects pretty much all day every day. Key informant interviews, bellwether interviews, document review, focus groups, you name it… So I know from personal experience that 99.9 percent of qualitative reports look like this:

    qualitative_before

    Which is fine, unless you want someone to, ya know, actually read your report.

    Here are several ideas for presenting qualitative data for your organization’s reports, presentations, handouts, infographics, and more.

    Word Clouds

    Data visualization novices love to love word clouds, while data visualization experts love to hate word clouds.

    Word Clouds Are Okay for Visualizing One-Word Descriptions, But Not for Visualizing Allllllll Your Qualitative Data

    Here’s the most obvious strategy for visualizing text-based data: the word cloud, also known as a tag cloud. Frequent words or phrases are shown in larger, bolder font. Less-frequent words or phrases are shown in a smaller font.

    I’m not advocating for run-of-the-mill word clouds where you simply dump your interview transcripts into a word cloud and hope for the best. Simply looking at how often a word or phrase appears in your dataset is not a sufficient way to analyze your data!

    Instead, I’m talking about intentional word clouds – when you’ve considered other options, and have purposefully chosen a word cloud as your visualization of choice.

    Here’s one example. When I conduct interviews as part of a qualitative research project, I often included a question in the interview protocol such as, “What one word you use to describe _____?” For example, “Mr. Policymaker, what one word would you use to describe public opinion towards poverty in the United States?” or “Mrs. Principal, what one word would you use to describe this new initiative’s effect, if any, on your charter school?”

    The rest of the interview is obviously focused on loooooooong responses, and fuller descriptions, and plenty of examples. But for one or two moments during the conversation, it’s nice to pause and throw in a semi-qualitative short question.

    Then, you’d visualize those one-word responses in a word cloud, like this one from Students First:

     

    People described their favorite teacher using only one word and the adjectives were visualized in a word cloud shaped like an apple.

    People described their favorite teacher using only one word and the adjectives were visualized in a word cloud shaped like an apple.

    Here’s another example from the Pew Research Center, in which they visualized one-word descriptions of Barack Obama with packed bubbles. Packed bubbles are similar to word clouds. In word clouds, word frequency is depicted through font size. In packed bubbles, word frequency is depicted through bubble size.

    Here's another example from the Pew Research Center, in which they visualized one-word descriptions of Barack Obama with packed bubbles. Packed bubbles are similar to word clouds. In word clouds, word frequency is depicted through font size. In packed bubbles, word frequency is depicted through bubble size.

    People described Barack Obama using only one word and the adjectives were visualized in a bubble cloud (and then color-coded by the sentiment or tone of that adjective).

     

    Word Clouds Are Okay for Visualizing Before/After Comparisons

    Word clouds are also great for before/after comparisons, like these tweets describing breakups. Does your study involve pre/post tests with a few open-ended questions? Did you interview participants at multiple intervals during the study? You could adapt this technique for nearly any time series design.

    The Best Software Tools for Creating Word Clouds

    There are tons of free and low-cost websites for generating word clouds. The Poll Everywhere blog summarized nine of their favorite tools. My personal favorites are Wordle and Tagxedo.

    Wordle

    Wordle is a free website for making word clouds in an oval or rectangular shape. It’s pretty easy to use. You simply copy all of your text (e.g., from your interview transcripts) and paste the text into Wordle. Then, you can customize the word cloud’s fonts, colors, and general shape to your liking. I suggest matching the fonts and colors to your project’s existing branding guidelines so that your word cloud matches the rest of your report or slideshow.

    Wordle is a free website for making word clouds in an oval or rectangular shape. It's pretty easy to use. You simply copy all of your text (e.g., from your interview transcripts) and paste the text into Wordle. Then, you can customize the word cloud's fonts, colors, and general shape to your liking. I suggest matching the fonts and colors to your project's existing branding guidelines so that your word cloud matches the rest of your report or slideshow.

    Tagxedo

    Tagxedo is a free website for making word clouds in fun shapes, like an apple, heart, or outline of the United States. You can pull in qualitative data directly from a website or Twitter feed (you don’t have to copy and paste, so you save a step). And, with Tagxedo, you could even have your word cloud printed on a t-shirt, tote bag, or coffee mug. I see Tagxedo as more of a decorative option than a researchy option.

    Tagxedo is a free website for making word clouds in fun shapes, like an apple, heart, or outline of the United States. You can pull in qualitative data directly from a website or Twitter feed (you don't have to copy and paste, so you save a step). And, with Tagxedo, you could even have your word cloud printed on a t-shirt, tote bag, or coffee mug. I see Tagxedo as more of a decorative option than a researchy option.

     

    Showcase Open-Ended Survey Data Beside Closed-Ended Data

    What’s better than quantitative data? Or better than qualitative data? Quantitative and qualitative data combined!

    Use this technique when your survey has both closed-ended and open-ended questions. Tie the responses together in one chart to add much-needed context. In this example, a survey asked nonprofits to describe what it was like to work with an outside consultant. Then, the survey asked them why their experiences were good or bad. Who cares if 33 percent of nonprofits had excellent experiences but we don’t have examples that describe why?

    Rather than simply listing out the open-ended responses in your appendix, showcase them beside your stacked column chart.

    qualitative_quotes

    Read more about this technique in this post written by Johanna Morariu and I.

     

    Include Photos Beside Participants’ Responses

    For your non-anonymous reporting, how about inserting photos of the interviewees next to their ideas?

    Here’s an example from the New York Times, where they added photographs beside interviewees’ responses:

    Here's an example from the New York Times, where they added photographs beside interviewees' responses.

     

    Include Icons Beside Descriptions and Responses

    Icons are so easy to use that there’s really no excuse for not using them to break up long sections of text.

    Another great use of icons comes from the Washington Post’s story about how animals are faring after the BP oil spill. Isn’t it remarkable how icons and bold subtitles help to break up a chunk of text?
    qual_icons

    The Center on Budget and Policy Priorities used icons to visualize how working-family tax credits can help at every stage of life:

    The Center on Budget and Policy Priorities used icons to visualize how working-family tax credits can help at every stage of life.

    And the Center on Budget and Policy Priorities used icons again to help their readers understand how low-wage workers face a cut in tax credits:

    The Center on Budget and Policy Priorities used icons again to help their readers understand how low-wage workers face a cut in tax credits.

     

    Create Diagrams to Explain Complex Concepts and Processes

    You can create diagrams for many aspects of your project, like this diagram explaining what type of protective gear is keeping doctors safe from Ebola:

     

    Graphic Timelines

    Regular text-based timelines + diagrams, photos, and other images = graphic timelines. Timelining is especially valuable in situations in which you need to track how a program, initiative, or campaign unfolds over time.

    Interactive Timelines

    This timeline from the New York Times includes an assortment of qualitative and quantitative information: dates, milestones, # of rockets, and # of casualties. My screenshot doesn’t do it justice; head over to the New York Times website to explore the interactive version.
    qual_timeline

    Static Timelines

    Maybe you’re not the New York Times, I get it. Anyone can design a static timeline from the comfort of their PowerPoint slide.

    Color-Coded Phrases

    One of my all-time favorite examples of qualitative data visualization comes from the New York Times’ election coverage in 2016. They compared and contrasted speeches from Donald Trump and Hillary Clinton. First, the New York Times team presented miniature thumbnail images of each nominee’s convention speech. You can even click on the thumbnail images and they’ll expand so that you can read the transcript. I love how the thumbnails provide a birds-eye-view analysis of the qualitative themes from each speech simply by color-coding certain phrases.

    One of my all-time favorite examples of qualitative data visualization comes from the New York Times' election coverage in 2016. They compared and contrasted speeches from Donald Trump and Hillary Clinton. First, the New York Times team presented miniature thumbnail images of each nominee's convention speech. You can even click on the thumbnail images and they'll expand so that you can read the transcript. I love how the thumbnails provide a birds-eye-view analysis of the qualitative themes from each speech simply by color-coding certain phrases.

    Directly underneath the thumbnails, the New York Times team pulled out a few sample quotes so that readers can get a sense of what was said.

    Directly underneath the thumbnails, the New York Times team pulled out a few sample quotes so that readers can get a sense of what was said.

    Want to try this technique yourself? Color-coded text is deceptively simple. I suggest highlighting the entire phrase in a light color, as the New York Times has done in this example. Make sure you don’t just color-coded the font itself. It’s challenging for our eyes to read green letters, but it’s pretty easy for our eyes to read black letters against light green background shading.

     

    Comment and Share Your Ideas

    How are you visualizing qualitative data interviews, focus groups, surveys, document reviews, and other qualitative data sources? Please link to your favorite resources below. Let’s give virtual high fives to the creators and celebrate a job well done.

     

    12 Comments

      Leave a Reply

      Your email address will not be published. Required fields are marked *

      Related Courses

      Most “professional” reports are too long, dense, and jargony. Transform your reports with these practical tips. You’ll never look at reports the same way again.

      Learn More

      Data visualization best practices, practical how-tos, tutorials in multiple software platforms, and guest experts. Designed with busy number-crunchers in mind.

      Learn More

      SPONSORED

      #f-post-el-35{display:none !important}

      #f-post-el-35{display:none !important}

      Learn More