I like data-driven storytelling (I like loaded words)

So these two things happened pretty much at the same time.

Alberto Cairo said this (and the conversation it sparked was absolutely superb) …

… and Jon Schwabish and Robert Kosara talked about data stories over at the PolicyViz podcast (which I have yet to listen to in its entirety because I want this essay to be a thinking-out-loud exercise —and I kind of already know where it might go.)

So … what’s up with that?


I woke up to the fact data storytelling was a loaded binomial when I saw my two favorite speakers from Malofiej 2016, Archie Tse and Kim Rees argue over it during lunch —I think Brian Boyer was the one who poked the bear. Archie was all for storytelling, Kim was all against it. I remember Kim got hilariously irritated by the word, especially when someone used storyteller to refer to her (that was me). (It was awesome to see someone other than myself getting overly worked up about nomenclature.)

That same year, I had the privilege of being invited to a seminar about Data-Driven Storytelling at Schloss Dagstuhl, where some really brilliant minds (myself excluded) deliberated about narrative techniques, exploration v. explanation, from analysis to communication, audience, ethics … At one point, as the conversation about topics was getting sidetracked, Nick Diakopoulos said something like “maybe we should start by narrowing the scope, what do we consider data-driven storytelling”. Oh boy! 😲💥🤕


But see, I think the problem isn’t in the definition of data-driven storytelling, or of storytelling, or of what we call a story, per se but where it fits in our process. To me, it almost always comes last, at the putting-things-together stage.

Here’re few reasons why I think storytelling in the context of data-driven journalism is important:

  • Too many pieces of data journalism fail to explain existing connections between the different bits of data it guides readers through.
  • The data we care about is an abstraction of people’s identities, experiences, and behaviors. Characters and their dramas are more relatable, bringing those out makes stories more memorable.
  • Understanding narrative structure, the relationship between the parts and the whole, makes for better editors.
  • Narrative templates come in many shapes. And many interactive devices do have a narrative structure. (I may need another post to explain this one, I’ll leave it at this for now.)

And here are the associated problems:

  • Storytelling can’t come before data, facts, evidence. The reporting, the research, the data analysis shapes the story. If you want the data to fit a narrative maybe you should consider a career in … (Nah, I won’t throw any other profession under the bus).
  • The drama of the characters becomes so fascinating that we fail to make the connection with the larger points of our stories, and the story becomes a collection of voices, no tie to an overarching structure.

Good data stories can either explain the cause-and-effect between phenomena, point to the possible reasons why things correlate or simply explain the reasons behind certain patterns.

In all these examples, storytelling, connecting the dots, happens hierarchically at different levels, between different pieces.


I’m not gonna delve into the benefits of building character-centric data stories. We know most of us care more about actual humans than their abstraction —the numbers. Best if you just watch this video and visual feature by The New York Times ‘A Bullet Could Hit Me and My Kids Anytime’. You know what I mean, right?

This is one of my favorite data-driven stories from last year.

Framing, choosing a narrative point of view, can also turn a seemingly exploratory device into a story. Take ‘Unaffordable country’, still one of my favorite collaborations between the Visuals desk and Helena Bengtsson’s Data team at The Guardian. It starts with a question, then set up, your choice, then the immediate result of your choice, then more detailed results. (Ben Shneiderman’s Visual Information Seeking Mantra can be thought of as a narrative template.)


So … again, what’s up with that?

Storytelling, especially factual, unbiased, data-driven, is so necessary to counteract propaganda. Educational researchers and cognitive scientists have long argued that narrative helps us organize our human experience. I get why it’s a loaded word, but we know it’s one of the best tools we have to make others understand the matrix we’re living in and fight disinformation.

So. Me? I don’t mind it. Kinda like it.

I guess it’s up to your journalism professor or your editor to make sure you don’t put it before your real purpose … producing ad space inventory and hooking enough users ... I mean, helping your audience understand the world and make informed decisions.