What Good is Dat?


Derek Willis


July 2, 2013

Max Ogden, a one-man band of interesting civic (and other) coding, has a new project called dat. Dat is “a new initiative that seeks to increase the traction of the open data movement by providing better tools for collaboration.” In practical terms, dat will consist of a set of tools to “store, synchronize, manipulate and collaborate in a decentralized fashion on sets of data, hopefully enabling platforms analogous to GitHub to be built on top of it.”

Great! More data tools! Or: Great. More data tools.

Github’s Ben Balter falls into the latter camp, asking why more tools for government data are needed: “It’s the culture that we’ve go to worry about. Will the solution be adopted by the agency? Will existing habits subvert the cool thing you just made? Will the dream live on once you’re at your next gig? If you ignore culture, even the most elegant of technological solutions can be relegated to a dusty shelf as nothing more than the right solution to the wrong problem.”

Ben has worked inside government and outside it. He’s not uninformed when it comes to this stuff. And he has a fairly accurate assessment of the whole government data scene:

I watch day in and day out as many of my former colleagues fight tooth-and-nail trying to convince well-meaning government bureaucrats to toss a scrap of government data over the firewall. It’s a tiring process. After all, whack-a-mole is, by definition, a loosing (sic) game. But the answer’s not Yet Another Mallet, nor is it to give up and build our own mole management solution. We should be making it dumb-simple to do the right thing. We should be building really, really boring stuff. The more boring the better. In many cases, we probably shouldn’t be building anything at all. This is one of them.

Max provides a response that addresses why git and Github don’t fit the bill for this project, but his response isn’t particularly relevant to my interests anyway. Ben’s right: the technology isn’t the issue. But that’s why I think dat is a good thing: because the best part of it, the thing that makes it exciting to me, is boring but essential and not really about the technology. It’s not the tools to store or synchronize the data. It’s not even the bit about transforming it into more useful formats.

It is the simple idea that if enough people use dat to explore government data, they will find the best and worst of it, document it and ultimately produce better things. This has been the case in my experience with campaign finance data, which requires solutions for storage, synchronization and transformation. Those of us who work with it have our own approaches, but the best things happen when we collaborate on our methodologies. We reduce the chance that any of us display or report figures that conflict with each other, and crucially, we learn more about the data in doing so.

That’s what I’m looking forward to in dat - a true platform for collaboration that enables users of government data to annotate it and to share their experiences in the hopes of creating a set of best practices for individual datasets. Maybe that won’t happen, or maybe the technical aspects will prove more popular than the collaborative ones. But the more opportunities for collaboration, the better, and I’m glad to see another one underway.