Be the Hemingway of data science storytelling

John Weathington
Updated

Churchill Downs in Louisville, KY

 Image: Lyndsey Gilpin/TechRepublic

churchilldowns2014.jpg

The right information in the right format has a tremendous power to influence even those who have little patience for analytics. A statement like, "50 people in the United States will die prematurely in the next hour, from smoking or exposure to secondhand smoke," will get anyone's attention. And many people who are on the fence of making a decision can be persuaded by the right data. However, it takes more than crunching numbers to move people.

Data scientists are often criticized for not knowing how to communicate to the masses, much less inspire them into action. You're not forced to inherit that script, but you should take some lessons from the greats to turn that narrative around. To tell a great story with data science, you must think like a great storyteller.

Know the format for a great story

About 75% of the population feels data are pretty dry and boring, but everyone loves a good story. Unfortunately, storytelling isn't an arrow you typically find in a data scientist's quiver, so you'll need some basics on how great stories work.

Stories aren't formulaic, but you'd be surprised at how consistent their general format is. A great story will always have a protagonist (i.e., the main character) who usually embarks on a quest or a journey. It's not necessarily a physical journey -- it could be an internal journey of growth or self-exploration. Along the journey, there will be challenges: dragons to slay or internal demons to overcome. In the end, there will be some sort of resolution, whether or not the protagonist succeeds (or even survives). Let's use this same general format to tell a great story with our analytics.

Bring your story to life

Think about the story you want to tell, and then use information, knowledge, and/or wisdom to bring the elements to life. Remember, information is the "so what?" of data; knowledge is the amalgamation of disparate pieces of information together for new and unexpected insights; and wisdom comes when information and knowledge is matured over time.

For example, a horse that runs a quarter-mile in 22 seconds is interesting data. The information is this is lightning fast compared to other horses; the knowledge comes from a dry weather forecast and a different trainer -- one who's more skilled at working with fast horses; and wisdom is knowing the track's history of catering to front-running horses, especially when the weather's dry. Sounds like a good bet to me. Let's apply this to the format of a great story.

Choose a protagonist, and then use information and the other evolutions of data to add depth to the character. I just did it in the example above with our fictional racehorse. Our horse isn't any ordinary horse -- this is a horse that runs a quarter-mile in 22 seconds, which is quite remarkable. You should use ordinary and extraordinary points of information to make your protagonist real and relatable, but also unique and interesting.

Next, set up and explain your journey. What are you trying to accomplish? What is your goal? What does your current situation look like? What challenges have you already faced and/or will face? You can bring all of this to life with information. I recently analyzed an organization's usage patterns with a particular process to understand how we might install better practices for the entire organization, and I found that only a handful of people are doing the bulk of the work. So, the journey for the organization is to move to an inter-dependent state, supported by its heavy users. But, I wouldn't be able to position my recommendation without the information to back it up.

Finally, there should be a resolution to your story. It could be an outcome that already happened or a future vision that you anticipate, but it should bring closure to your story. Do not start your target audience down a path that doesn't lead to anywhere. It doesn't have to be a happy ending, if your point is to influence people in the opposite direction. I often tell the story of the Bhopal, India disaster in 1984 that killed 2,259 people in three days. I use the information around this tragedy to highlight the importance of process safety. Notice how specific I am with the data. This makes the story powerful -- if you follow the right format.

Summary

Telling a great story with data is not hard if you follow the lead of the masters. Like Hemingway, Fitzgerald, and many other great novelists, craft a story with a protagonist, a quest, and a resolution. Then, add dimension to your story with information that drives a point, knowledge that brings amazing insights, and wisdom that cradles the intended audience in the comfort of lessons past learned.

Data science may be dry and boring to your target market, but they don't need to know that! Package your analysis within a great story, and carefully guide your audience to where they need to be.

Also see

Note: TechRepublic, CNET, and ZDNet are CBS Interactive properties.