Six months into her new position as New York City’s Chief Analytics Officer and Director of the Mayor’s Office of Data Analytics (MODA), Kelly Jin has sustainability on her mind—technological sustainability. Tasked with leading data-driven governance across all city agencies and the open data portal of the largest metropolitan area in the world by urban landmass, Jin and her team of eight must work with what they’ve got—which means working smarter.
Luckily, Jin brings to the table experience in data analytics at both the federal and city levels—from the White House and Boston City Hall—and from the private sector at the Laura and John Arnold Foundation, where she served as the Director of Results-Driven Government. At MODA, Jin must maneuver through government bureaucracy to both bring the city’s technology into the twenty-first century, and identify where city governance can be more proficient with data analytics.
In an interview conducted at the Harvard Kennedy School of Government, Jin outlined her goals in her new role, the projects MODA is currently working on, and her vision of how to get NYC’s approximately 50 city departments working more collaboratively and efficiently with data.
Stefanie Le: What are your goals in this role and how will you go about achieving them?
Kelly Jin: The first one is really focused on lifting the floor—how can we have the right level of training for data analysts across the city? One thing to know about the New York City environment is we have analytics teams that are distributed throughout different city agencies. Because MODA is a team of only eight, it’s really important to think about how we can provide resources, training, even technical expertise in support of those data analysts throughout the city.
I think more broadly beyond the fact that we’ve been written into the city charter, we need to figure out what all of the additional data-related needs are—not just analytics—throughout the city. That includes how we can improve how we do technology procurement when it comes to data systems, [and] understanding that analytics is only as good as the input data and data quality that we do have. Additionally, city agencies are still working with a lot of legacy systems that have a lot of that data on lockdown. What’s really important to me is to help solve for that.
SL: Have you come across any challenges so far in this job?
KJ: I think the challenges within MODA are similar to many challenges that other folks in other communities may face. For New York City specifically, it’s the scale. I think that that’s something that we continue to work through as a team. What value do we provide? How do we provide that value within a very, very large organization? I say this a lot—there’s only eight staff within MODA and that works itself out to about one MODA staff member per one million New Yorkers. Really, really being thoughtful around what we can do versus what we can’t do as a team is very important to me.
SL: Are there any MODA projects you’re excited about rolling out or that are currently happening?
KJ: For me, it is the next phase of the tenant anti-harassment work that we’ve been doing since 2015, when there was a tenant anti-harassment task force built up. MODA was tasked with working with a few different city agencies to integrate data and then helping assist city agencies to identify who the bad landlords are across the city. There was actually an announcement today of a new director of the Mayor’s Office [Jackie Bray] to protect tenants. So this is a really neat story of what the ‘version two’ of the work looks like when we are three, four, five years later. Furthermore, it all has to do with: what is the institutionalization and sustainability of the work? So it’s not just MODA doing a one-off project, but saying how can we make this sustainable? What are the technology procedures that we need to put into place?
SL: How has your previous work experience in Boston, at the White House, and the Arnold Foundation shaped your approach to MODA in New York?
KJ: I think going into any local government role, having additional context of how other agencies operate at different levels is helpful for background. So understanding some of the other policy domains that I’ve had experience in working with in Boston, specifically public safety and a bit in housing, and at the White House and the Arnold Foundation in criminal justice. Just literally providing some additional context when interacting with the city agencies in New York is critically important.
I think the other benefit is always having comparison points. And so saying, “Hey, this is how another city agency or another jurisdiction has approached this type of problem” and having other analogues is always critical so you’re never feeling like you’re recreating the wheel in any way, but that you have some best practices that you may be able to call on.
Whether you’re in New York City or you’re in Boston or you’re working at a foundation, the common thread is “how can we improve people’s lives?” In New York City, specifically, it’s “how can we make life more equitable and more fair for New Yorkers?” It’s important to me to continue to work within teams that have that as a guiding north star in the work. And I think that seeing how different institutions strive to do that in different ways means that when I’ve currently landed in New York City as Chief Analytics Officer, I’ve had the frame of seeing and fully assessing how the city is already doing what they’re doing and what value I can add to that.
Over the last six months I’ve spent time really listening and learning from folks and being the one oftentimes asking the questions of existing practices just to learn more. Because I think the flip of this is that while I’ve learned a lot in my previous jobs, and am applying those lessons in New York City,I’ve also realized that every organization you go to is different and not to try to solve a similar type problem in one city just because a solution may have worked in another jurisdiction.
SL: You also previously mentioned the Open Data Advisory Council—can you tell me more about that?
KJ: The Open Data Advisory Council was formally selected in early March and we have 23 different open data advisory council members from all five boroughs in New York City. Our vision here is, “how can we bring in the broader community to help us set what the vision is for open data moving forward for the city?” Traditionally, we do an Open Data Report each year—it’s an annual report that we’re required to do as a city that has traditionally been a compliance. We are seven years out from the Open Data Law, and Adrienne Schmoeker, who’s the Deputy Chief Analytics Officer, and I really wanted to put our heads together and say, “what does the future look like?” We didn’t want to just do that being MODA or being the city--we wanted to open that up. So I’m really excited—we’re going to have a lot of different voices, some of whom have always served as informal advisors to us throughout the years. They’ll be helping us guide what the open data report itself looks like, which we’ll release in September. And then, they’ll convene twice a year. We have a lot of heads of New York City organizations, many of them users of open data, and that’s so critically important. It’s not just open data as a policy but really tactically, what are some improvements we can make for the public?
SL: What’s the Open Data Law that you mentioned?
KJ: The Open Data Law in 2012 in New York City formalizes a lot of what you actually see in practice today, which includes the fact that city agencies must have open data coordinators. They also must report on all of their data that is publicly available and then they must put all of that in a machine-readable format now through the New York City Open Data Portal. Because it’s been written into law, now there’s an annual compliance mechanism where we are working with city agencies. There’s also a fair amount of educating and training because city agencies—depending on their size—have different numbers of open datasets. We have over 2,300 datasets on the Open Data Portal. Additionally, we need to better understand what the public is really using on the portal, and a lot of that has to come down to context. How are city agencies providing more context behind the data that’s actually going on the website?
Because the data is operational data, it’s information city agencies are collecting as they are going about their business and doing their work. Explaining what all of that means, whether that’s as granular as what a column field actually means or how they’ve set a date/time field—that’s really important because it helps to translate that to the public or let’s say a nonprofit or even a student who’s doing a research assignment so they can actually derive value out of an open dataset.
SL: How do you decipher between what is and isn’t an analytics report or MODA project?
KJ: I think that more broadly what it means to be an analytics question or a problem is a constant education process. Because when you work with a city agency, the initial ask may not actually be a clearly defined analytics problem. And so the first question often is—does MODA have access to this data? We may respond: why are you asking for this data? And there may be a response saying, “Oh we’re trying to figure out more information about this particular business problem that we’re having.” When you define an analytics problem for MODA specifically, it always comes down to what’s an operational change that a city agency could make to help improve their service delivery or improve their value to New Yorkers or to the city. But when you open with that frame, that oftentimes will not directly translate into a beautifully scoped document from a city agency. And so we find a lot of our critically informed work is at the start—sitting down with folks or on the phone going through what their needs are, and helping them refine what their analytics problem statement truly is. One thing I always say about the team is we always pick up the phone and we understand that city agencies are looking for help and they’re looking for resources. We definitely want to have that conversation and if it’s not a MODA project in this moment, say “here are some additional things that you all can go look at in terms of collecting additional information or data or here’s some other resources that we can connect you to.” Again, I think one of our limiting factors is there are only four data scientists and eight MODA staff members. The team definitely punches above its weight, but we need to be mindful that you can only move forward on so many long-term projects or short-sprints at any given point in time to run a successful, sustainable team.
SL: When did you first realize the importance of data analytics for city governance?
KJ: I would say in Boston. Mayor Walsh had just been elected and Dan Koh was his chief of staff at the time. Dan and the mayor really wanted to figure out: how can we make a more data-driven city of Boston? And my time spent at the federal level was around culture change using data and technology, so it was really interesting and appealing to me. I had gotten a call to see if I might be interested in helping to stand up the data team at the outset. At the time, I was leading up data visualization for the city, but I think what’s always been really, really important to me is that it’s not just analytics for analytics sake. It’s about how are we actually making an impact on the ground and how are we changing action or operations of a frontline worker or of a frontline manager because that’s actually where government happens. Where a government employee may interact with a member of the public—as resident or a visitor in any particular city. And so I had seen culture change happen through data and tech at the federal level and was very keen on exploring what that might look like at the local level and at City Hall in Boston. Clearly, I couldn’t stay away because I’ve now come full circle and gone to New York City because to me it’s that service, that on-the-ground feedback of what is working and isn’t working. There’s nothing like working at the local government to actually see that feedback cycle in a 24-hour or even less kind of span.
This interview has been edited for clarity and length.