Data for Good Exchange (Bloomberg)

Last weekend, I went to New York to attend Bloomberg’s Data for Good Exchange (and also hang out with my friends, eat bagels, and have brunch - obviously). I came in with pretty low expectations and left with a lot of excitement for finding myself a way into this field.

When I started grad school, when people asked me what I wanted to do after my PhD the only answers I had were things I didn’t want to do (academia, industry, policy, high-tech bioengineering start up). Then, I started figuring out the individual parts more clearly: I want to work in a field related to health, I want to use my technical skills (i.e. brainz and computerz), and I want my work to disproportionately impact underserved populations. Okay, that was cool to figure out but I still didn’t really know what actual jobs were at the center of that Venn diagram.

So when I found the Data for Good community, I got really excited! I had a vague idea that public health was a place where my technical skills could be useful and where I could maybe have the kind of impact that I wanted - but I didn’t know whether my frame of thinking was already valued in that space or if I’d have to try really hard to convince the people with the jobs that I have the intersection of skills and interests that they want. So to find a community of people specifically interested in applying their computational skills to advancing the social good - so awesome!

That said, I’m really wary when I see the words “technologists” and “social good” in the same paragraph. It just brings me right back to the discomfort I had with the NGO “scene” in Cambodia. I’d been following a few blogs and had a pretty good feeling about it, but I still wanted to see for myself how much hype vs. content the actual people in the community have. The Data for Good Exchange seemed like a great place to start figuring this out, so I went into it not really expecting much. 

At first, I walked into Bloomberg and was blown away by the swankiness of the place - beautiful, open spaces and free snacks as far as the eye could see. Okay, still skeptical. Then I walked into the main meeting room and saw tote bags and useless swag - uh oh. (How much “Social Good” could Bloomberg have done with the dollars they spent on the wood puzzles they gave us?). The first keynote had a hype-to-content ratio of like 8 to 3 - telling us that opaque machine learning algorithms that de-humanize people are bad, both ethically and  technically. Yeah, duh.

But then the conference really got rolling. The second keynote was a panel with the White House Office of Science and Technology Policy. Awesome. And then almost every single talk I saw for the rest of the day had both the data scientists who did the data work and their on-the-ground partner - non-profits, foundations, public/governmental organizations. And what’s more, it’s clear that the projects were truly partnerships - it wasn’t like the partner gave their spiel and then the data scientist gave theirs and neither one knew about what the other was talking about, the presenters clearly knew each other well and were engaged in a continuing conversation. These projects were not disconnected from reality and then loosely tied back with a “Broader Impact” fluffy fluff statement (like is so common in academia) - they were really projects born of true partnerships. Which is so great - sign me up! :D

The coolest part of the conference was the “Future of Data Work at Aid Agencies” panel. The panelists were “chief something something” at World Bank, UNICEF, and USAID (plus someone from Facebook working on public policy something something). Dayum. The conversation was also really enlightening, and definitely made me consider a new way of looking at global aid work. Yes, I have issues with how neo-colonialist aid is currently approached, but those issues maybe apply less at the higher levels of these large aid organizations. For example, compiling GDP statistics or figuring out better ways to collect data to monitor impact - this kind of work doesn’t have to be done by local people with local context, and is more, like, just a job. In other words, there’s lots of different ways to get involved with aid work - and not all of it is riddled with the complexity of white savior complex etc.

Something else that struck me from the aid panel was the different timescales of data. I think a lot of people think that more and faster data is always better. But the guys at these aid organizations were like, meh - sometimes it’s better but sometimes you really only need one survey every five years (for example, to measure the impact of an education initiative). That said, data-informed change can happen much faster than organizational structures often let it - and so the conflict between this fast-paced world that we live in now and the necessarily slow pace of systemic change is really fascinating. I loved the maturity with which the panel approached data work in aid - they recognized data’s value and power but didn’t see these large organizations as limiting data’s impact. In fact, even though working within these massive organizations is slow and blundering, it’s also incredibly powerful and has the potential to turn new best practices into worldwide adopted standards. If that’s not impact, I don’t know what is.

Related to this conflict between big, slow structures and fast, weasely data was the idea of private vs. public work and money. One of the speakers from the University of Chicago’s Crime Lab said that their goal was to “use private dollars to leverage the public dollars.” In other words, they fund studies to gather data on the impact of existing small initiatives so that they can show the results to government officials and convince them to scale up the initiatives to the city- or state-level. Again, I really love this attitude of respecting the kind of systemic change government is able to make - if you just give it the chance and the power to tell its opponents, “No, I’m definitely sure this will lead to positive change.” That said, the people we saw at the conference are already data-literate and data-sympathetic. How do we convince the others who aren’t at this conference? I’d love to see conferences like this explicitly address the communication gap that I’m sure exists between scientists and public officials. Analyses are nothing if they’re not paid attention to.

Essentially, my goal in going to this conference was to figure out if I should keep going to these conferences. From what I saw this weekend, I think the answer is a resounding yes. So excited to have found a community of people who are so motivated to do good with good science.