Building a Better City Through Big Data

In the nation’s capital, 28 percent of children live in a household that’s below the federal poverty line, and another 20 percent grow up barely above it. As executive director of DC Action for Children, NationSwell Council member HyeSook Chung studied exactly where this deprivation could be found and, more importantly, why. “What are we doing that’s not working, and why are we investing in it?” she asks repeatedly. Unlike the ideological think tanks that populate D.C.’s corridors, she’s a relentless empiricist who searches for answers in data. At DC Action, she partnered with DataKind and joined the Annie E. Casey Foundation’s Kids Count community to publicly post a number of visuals about the city online, graphically comparing, say, youth unemployment, Medicaid enrollment or the number of parks in every D.C. neighborhood. Last month, Chung accepted a new role as D.C.’s deputy mayor for health and human services. As she makes the transition, NationSwell caught up with her to discuss the data-driven accomplishments at her last job and reflect on what her new role means for the city.
How does better data guide decision-making in Washington, D.C.?
At DC Action, we were the first ones to really look at the neighborhood level. Looking at wards — the equivalent of a county level — was too broad. As a parent, I live in D.C. and my kids go to DCPS, and I wanted to know why parents in certain areas were able to move the needle, despite the lack of support from the city’s administrative offices. With neighborhood data, we could question why a cluster of a few elementary schools were doing better than all the others in that ward. It could be race or income, but I wanted to know exactly why.
That led to visual analysis and asset-mapping that we can show a council member. “Look at grocery stores and the lack of fresh produce in Wards 7 and 8. Look at the poverty in Wards 1, 4 and 5 that’s starting to kick up.” We were able to have a different conversation with city leaders. Some of the big fights in the city are about state representation and all the things happening on the Hill, so I don’t think they were ready for an organization to show up with data on the neighborhood level. Because then, the solutions are really localized solutions, not these macro, citywide policies. That’s a different way of thinking: One solution is not going to meet the needs of all 108,000 kids under 18.
There’s been a lot of debate about how data can be misused. How do you avoid trusting misleading figures or building biased algorithms?
Data is not so black and white, especially in human resources. People dealing with people is very subjective. How can you have an automated evaluation for hiring or firing? In public education, there’s this drive for outcomes in test scores that need to be improved if the teacher is to be effective. I heard from one teacher who scored 6 percent [in his evaluations] one year, then 97 percent the next. The educator said that nothing changed; the calculations were just different those two times. Their salaries, pensions, even their jobs are determined by these equations some person is putting together. That is one thing about open data about which we have to be conscientious.
As the repository for Kids Count at DC Action, we focused on making sure we had the most up-to-date, reliable, unbiased data out there, but we also kept track of how that data is used. We all have biases that data can further or can debunk. We took our role very seriously to be as unbiased as we could, to give as much context as we could, then let the data speak for itself.
How can service providers change their operations to keep better track of their data?
I was training a few of the intake coordinators at one community-based organization, and I walked them through why everything they do is so important to track. I referenced Amazon: As a user, every movement, every click is tracked to give me popups based on what I might like. For nonprofits, the only difference is you meet families and children every day, and you have all these interactions and conversations. But none of that is being recorded or tracked. One of the pitfalls of social finance data is that we’re very great about tracking quantities and caseloads, like how many families you served or how many kids graduated, but we’re not so good about tracking progress or the quality of services. That’s been something I’ve been pushing recently: It shouldn’t be about how many preschool slots we have, because we have to narrow down how many of those are quality. They’re not all equal. We’re trying to set a new bar. Caseloads are not enough information to show progress.

DC Action, in making public data widely available, is really just scratching the surface on the reams of information agencies could collect. What does the future look like if the public sector fully embraces this tool?
Can you imagine what the impact would be on the social-service sector if we had real-time data? It’s profound: Netflix and Amazon are able to adjust, in a matter of seconds, based on consumer knowledge. At nonprofits, we have a long way to go to embrace that and redefine accountability. Of course, it’s not truly transferrable from the private sector, but our decisions about service delivery could be much more engaged and responsive to live information from a family. We have to be careful; we don’t want to profile. But how do we translate, with these ethical and business questions in mind, those insights to the social sector to be more effective for families? That’s my interest. I want to get to a place where we can say, “Because of this investment here, we had this result.” It’s not about money; it’s about how we use the resources we have. If a program is not improving outcomes, have the courage and the data to adapt it. We’re not quick enough, and that’s frustrating to me. I just don’t know why we are in this rut of not giving our kids what they deserve.
How do you define leadership?
Two words come to mind: integrity and resiliency. Being an executive director is really hard work. I’ve made decisions, I’ve dealt with funding changes, I’ve let go of friends and fired people. At the end of the day, if my integrity is intact, I can go to bed, knowing I did the best I could. There were plenty of times I cried a lot and had to make hard decisions. But the work continues, because the bottom line is kids need us. The mission keeps us moving.
Why did you decided to take a new job in city administration?
At DC Action, we were called upon by the mayor’s executive offices to help make data-informed decisions. In many ways, we were partners in an advisory capacity helping departments achieve results and made decisions based on outcomes, not simply compliance. After meeting with Mayor Muriel Bowser, I knew [this job] was another wonderful opportunity to push our starting principles to a much larger scale. The mayor invited me into the administration to help highlight the critical importance of data-driven work for some of the toughest challenges we have before us as a city: homelessness and reform of the Temporary Assistance for Needy Families benefits.  As a public servant, I am thrilled to be asked to think more strategically and systematically about how we can truly make a difference in the lives of our residents in need.
To learn more about the NationSwell Council, click here.