I’ve been a fan of macabre fiction writer Stephen King since I first picked up The Dark Half sometime in the early 1990s. Since then I’ve read dozens of his books, and I’ve never been disappointed by a single one. However, when I finally got around to reading the two-decades-old book On Writing, I found not only some incredibly helpful writing tips, but a lesson in data storytelling.
In the pages of On Writing, King writes a nonfiction account of his own development as a writer and openly shares some of the methods he uses when penning a manuscript. Since my writing is almost exclusively technical, many of the nonfiction-specific tips weren’t specifically helpful to me. However, when he shared about how a story normally develops in his mind, he surprised me with the way he approaches the evolution of a tale:
I often have an idea what the outcome may be, but I have never demanded of a set of characters that they do things my way. On the contrary, I want them to do things their way. In some instances, the outcome is what I visualized. In most, however, it’s something I never expected.
I’ve written maybe 20 fiction stories in my life, always in an academic setting and rarely as a voluntary exercise. Each time I did this, I had the ending and most of the story complete in my mind before putting pen to paper. However, King’s description of his story development is the opposite of that: he starts with a fairly simple situation and transcribes the story as it develops in his mind, and the plot and the (often unpredictable) ending play out through the telling of the story.
Although fiction writing is wholly different than technical writing and blogging, I was struck by reading this at how telling a story with data shares some of the same mechanics. The process of taming raw data should not have a fixed plot and predictable ending with which the details (data) are massaged to support. Data isn’t the flair around an existing story: the data is the story. This applies whether we’re wrangling data from untrusted sources, building key performance indicators, or constructing exception reports to show what are perceived to be outliers.
We as data professionals should take heed of Stephen King’s direction and follow the story told by the data. Even if it surprises, offends, or contradicts our expectations. Especially if it does those things.