The problem with being “data-driven”

Data data data. Big data. Granular data. You won’t need gasoline, or a Tesla – you are data-driven.

It’s hard to get away from “data” (or its cousin, “information”). What does the data say? Does the information tell us anything?

I don’t know; I didn’t realize that these things were sentient beings that can “say” phrases and “tell” us about ideas. And their vagueness – what IS data? Does it have important physical characteristics or is it part of the world’s growing layers of abstraction, under which manual work and tangible items are obscured and devalued?

The issue with the world’s data obsession is not that it necessarily produces bad commentary, bad writing, or bad sociological analysis. Still, it does do that. The data told us that, without any Entente/American troops on German soil in 1918, Germany maybe shouldn’t have surrendered. Who knows what their odds of victory were?

It’s not all bad news. Being data-driven leads to big, nice-looking, slow-loading webpages such as Vox, The Verge, and FiveThirtyEight, and their newsrooms full of tone-deaf white guys.

But trusting data over all else is to shirk social responsibility. It’s to wring one’s hand in faux-seriousness to some intangible ideal – data, numeracy, whatever you want to label it – that really plays a role analogous to god (but that would be frowned upon by data-driven crowds, natch). It becomes the agent, something unstoppable and immutable, while the writer becomes less accountable – after all, she’s just letting the data talk, as if channeling the Burning bush.

What a naive viewpoint. Using data is a matter of interpretation, and many writers don’t have the chops for it. It’s no coincidence that two good pieces (one from The New Republic, another from Quartz) bemoaning the rise of data journalism were published on the same day (today). They point out this gap, plus they get at a bigger issue – that even data-driven writing is opinion, with research structured to favor a particular point. Problematic literature is sidebarred or ignored altogether – it basically has to be, whether the writer intends to do so or not, given the sheer volume of material out there.

It’s easy to see why data-driven writing has cachet on the Internet, with its somewhat technologist demographics. And maybe if FiveThirtyEight gets this year’s NCAA brackets right, it’ll have done a good deed – I’m not saying data journalists aren’t making valuable contributions. But their relentless drive toward “the future” of journalism or “journalism for the digital era” (how long have these trope been around? They’re fucking exhausting), like all progressivism, is often overzealous and blinkered.

I mean, look at education. What has decades of data-driven teaching, testing, and planning done? Billions in profits for private corporations, miserable students with their mouths taped shut, unions busted up, laughable Common Core standards…administrators have spent the past 30 years trying to close an “achievement gap” that doesn’t exist, ignoring poverty in their drive to throw data-driven strategies at things such as SAT testing that have little redeeming social value.

What does that show? Well, data itself isn’t important. It’s what people do with it, how they interpret it. How could one not see the broader U.S. trend toward privatization and inequality in the data-driven education craze, except with actions sanctioned by the authority of “the data says ___”? I worry that writing is going to become like this too, with terrible data-driven drivel designed for machines and condoned by the godlike vehicle of “technology.”

Blog at

%d bloggers like this: