When asked to create a Think Like a Data Storyteller workshop for SAS Global Forum, I realized that even if I had been telling data stories maybe I didn't really know what I was doing! After hours of research, I understood a million ways to tell a better data story but still didn't know if it the data story's purpose was to inform or persuade. I needed a definition.
Maybe Data Storytelling Is An Abyss
There were several definitions offered by various experts. All the definitions agreed it was a process of visualizing the results of analyzed data. Some of the definitions included words like "influence the business" or "determine the actions" to indicate that the story required an element of persuasion. Others were happy to stop at "telling someone else what you learned" point.
When you look for examples, it gets even more curious. I won't name names - but here's some of what I discovered.
- One article suggested a well-constructed chart with a descriptive title was a true data story. If the chart is Minard's Map of Napoleon's Russian Campaign, the example might be true.
- One promising article offered multiple ways to tell a data story but failed to deliver. It illustrated many charts. It suggested that a line chart was a just data story about changing values over time. There was a chart for the negative and positive trends because, of course, those are two different data stories. (Or was this just a product demo with a click-bait title?)
- One article suggested an awesome example was a video from Google that showed a query being typed, some swanky images, and a hip beat. No data was presented but the message "Yo! Google is in love with itself and does queries" was clear. Not a revealing plot line and not a data story.
Maybe There are Data Story Genres?
After completing my research, I see two data story genres -- informative and persuasive. Data journalists tell informative stories and marketers tell persuasive stories.
An informative data story is the Mapping New York's Noisiest Neighborhoods article by Ben Wellington in The New Yorker. This in-depth data story weaves a narrative through several charts. The surprising fact was that certain neighborhoods complained more about ice cream trucks. That is funny but not compelling. "Big cities are noisy" doesn't seem surprising. Wellington was explaining why they are noisy. His conclusion was "someone should solve this problem." So, he was given a platform, but he didn't make any change proposals. It felt like he wasted my time.
It's difficult to find persuasive data stories. The late Hans Rosling comes to mind. His data stories persuade viewers to make the world a better place. In his Numbers are Boring, People are Interesting talk, he uses himself as an example to encourage we change how we think about families. (His title gives away one of the secrets of good storytelling -- it's about people, not statistics.)
Here's My Definition
If data professionals want to bring data storytelling into the business world, the marketer's viewpoint is most useful. Many times, I have presented data where it was just explaining what happened without any focus or even a recommendation -- it was journalism. Like Professor Wellington, I had a platform but not a point. The times I have used data to prove a point and asked for change, I was the most effective.
Data professionals need to bring these skills to their work. Organizations spend a lot of effort organizing their data. The ability to turn the data into stories that can lead change will be essential in the coming years.
Here's my definition: "Data stories combine compelling data, valuable visualizations, and a strong narrative in a way that encourages organizational action."