Last night I attended the kick off for the Seattle Data for Democracy hackathon. With the proliferation of terrific open data sources such as the Seattle Open Data site, I was looking forward to meeting fellow politically-enthused data practitioners. The chance to employ data-driven exploration, analysis, and communication at a larger scale than my own one-off projects is exciting. Unfortunately, what I experienced were well-intentioned folks falling into the trap of Morozov’s technological solutionism and had to pull the eject lever once my fears were confirmed.

Scope (aka “Why are we here?”)

Data for Democracy (D4D) started in December of last year as part of the surge of public interest in the wake of the 2016 US presidential election. Many people, myself included, have been looking for ways to get engaged in their communities and politics. D4D lists a goal to “contribute to positive social change” but never defines what positive social change looks like. Being concrete on objectives is critical for any sort of data analysis project. In the absence of a defined scope, everything is fair game. Covering local traffic issues, international trade, tax policies, and beyond are all equally valuable. Far from being liberating, when everything is open, there’s no focus to your efforts. Without a goal in mind, progress is impossible. The local D4D event compounded this issue with an introductory conversation that refused to state advocacy for any particular position or theme. In liberal Seattle, if you’re unable to state your goals as either opposition to the Orange One’s administration or advocacy for particular progressive policies, then you’re expressing a timidity that is doomed to failure.

Audience (aka “Who are we trying to reach?”)

A data analysis project needs a clear sense of it’s intended audience. A project focused on board members making an investment decision has a very different feel from one performing investigative journalism for the general public of a city. Without understanding whom you are trying to reach, you’re denying yourself the opportunity to choose the best form and content type to provide clear information and responsibly influence decisions. In the case of the D4D event, the general directions were to spend five minutes discussing what positive social change would look like (this part only after my prompting) and then come up with project ideas within the next twenty minutes. No other guidance as to the audience was available. It’s still unclear to me whom the D4D project expects to consume the output of any of these projects. Sites such as the NY Times, FiveThirtyEight, and so forth may have a core of known consumers, but how the organizers expect anyone to find the results of these hackathon projects is beyond me.

Measuring Success (aka “How do we know when we’re done?”)

One of the most dangerous failings of this event was a lack of clarity, or even addressing, what success looks like. While impossible given the previously mentioned vagueness of goals, not having any way to measure impact of your work risks sapping the energy of the participants. Last night there were thirty or so folks being asked to give up an entire weekend of time to work on projects that likely will never reach anyone or illuminate a single decision. If you take a rough burden rate of $100 an hour across the weekend, that’s $93,000 of donated labor. What are the outcomes of that investment? Who is being reached and in what fashion? Most importantly, how disillusioned will first time participants of this event be if their sole exposure to data advocacy is doing some never-seen and never-used random python or JavaScript project? Code is the a means to an end, not the end itself.

Wrap Up

I may sound a bit like the hothead paisan and, make no mistake, I am more than a little angry what this missed opportunity means for the larger possibility of fact-based progressive democracy. There’s no shortage of ways to get engaged in random data science projects and I applaud any organization that wants to connect folks that want to work together, even if it’s just to fool around with data. Let’s not delude ourselves into thinking this is going to create world- (or even local-) changing effects. In the post-fact reality in which we find ourselves, the sort of finger painting dilettantism from the D4D group may be enjoyable and creative, but it’s going to produce neither meaningful value nor constructive dialog – the two things I seek most.