#MakeoverMonday can be incredibly frustrating, which is surprisingly a good thing. (More on this later.) Writers often grapple with writer’s block, but an adequate term that captures the struggles of a data viz professional is more elusive. This particular situation occurred during week 14’s #MakeoverMonday data set about the potential impacts of automation on the U.K. workforce.
I was initially excited about the automation data set #MakeoverMonday topic when Eva Murray announced it. Debates and discussions about how automation will impact the global economy and workforce are always interesting to me.
— David A Krupp (@DavidAKrupp) April 2, 2017
However, working with the limitations in the data quickly deflated my enthusiasm. Luckily, I saw a hilarious tweet by @mikevizneros, which let me know I wasn’t alone in this struggle. I actually laughed out loud when I saw it, because it captured my exasperation perfectly.
Live video of me trying to make something interesting for the “robots taking our jobs” #MakeoverMonday pic.twitter.com/72d2UgyAm1
— Miguel Cisneros (@mikevizneros) April 2, 2017
The limitations were not from data quality issues, but rather from not easily finding a good story. With only percentages of employment share versus percentages of automation risk, every insight seemed a little too obvious. Create a Gantt chart or bar chart by industry to compare the employment share versus automation. Easy peasy, let’s call it a day.
Joking aside, it was very difficult to think of a way to improve the original Gantt chart displayed in the article by PricewaterhouseCoopers. Grudgingly cycling through the usual chart options, I was not encouraged that I would be nearing a eureka moment.
How can I find something interesting? Knowing the other #MakeoverMonday participants would, like vultures, pick this simple data set clean of stories made the task especially challenging. Moving forward with a compelling story or a creative idea seemed impossible.
So I gave up and decided to quit. Week 14 would be skipped. Oh well. Mildly irritated, I rested on my bed. My next goal was to completely forget this episode and to begin thinking about something more relaxing.
Listening to music and staring at my ceiling, I eventually became at peace with my failure. After closing my eyes to begin a much needed nap, my missing inspiration finally struck. My thoughts screamed, Unit chart! I quickly sat up with energy and glanced at my computer screen from across the room.
While still sitting up in bed, I started visualizing different style choices. Finally, with direction, I began making key design decisions.
First, a giant 1,000-unit chart broken up by industry. Next, the unit shapes would be circles (not squares). The color selection would incorporate red, black, white, and grey with the goal of a New York Times graphics feel. Lastly, white/grey space would be exaggerated to avoid overwhelming the viewer.
This detailed design approach flourished in my mind within only a few minutes. It didn’t require sketching ideas on paper or tinkering around in Tableau. Reflecting on this specific experience still prompts me to wonder how it happened like this — mysteries around unpredictable idea generation have always fascinated me.
Having the design solidified, I still needed a captivating narrative. Luckily for me, unit charts are great for communicating ratios of whole objects, like people. After converting percentages to ratios, the data quickly transformed into something more tangible and thus more relatable to the audience. It’s much easier for people to relate to x number of people losing their jobs versus a percentage of people losing their jobs to automation.
Ratios also helped make the side bar chart relevant to the story because they showed how automation risk from each industry contributed to the U.K. overall. By what I still consider a fluke, all of the dashboard elements seemed to fit together perfectly like a puzzle.
Segmenting out the unit chart by industry also helped the story because the viewer could now answer multiple questions simultaneously: To what degree would automation impact the U.K. overall? How much would automation affect each individual industry? How important is each industry to workers in the U.K.? By adding context and combining these questions, it makes for a more compelling story about the potential impact of automation.
My next task focused on final touches. First, a red-colored robot image, which I designed in Adobe Illustrator, focused attention on at-risk jobs. Second, my original headline needed work. The first version was a flop because it didn’t explain the ratio numbers. Not realizing my oversight until I published my submission on Twitter, I quickly deleted the tweet before too many people could see my initial post. Oops!
I eventually updated the title to, “If Britain’s total workforce equaled 1,000 people, then 300 workers would be at risk of automation in 15 years.” This acted as both the main story and the explanation for the rest of the dashboard. My submission made the weekly roundup, which helped me reexamine the decisions I made throughout the process.
Every week in #MakeoverMonday varies slightly from prior weeks. New data sets present us with unpredictable challenges and opportunities. It’s the frustrating experiences, like the ones I described in this post, that we seem to learn from the most.
Do you have a similar story about how you came up with a good idea? Share it with me on Twitter or in the comments below.