Governance Matters: How did you come to be the Chief Data Officer of the City of San José and what does it entail?

Christine Keung: I would say I had a very non-traditional path to government. I joined the City of San José through a programme from Harvard Business School called the Leadership Fellows Program, which empowers people with a management and operations background to go and take high impact roles in leadership.

I think the folks who fund and support this programme really believe that beyond policy, government is also about service delivery. I accepted this fellowship in 2019, not expecting to serve it through a global pandemic, and so I sometimes do reflect on what my job would have been like if we were living in normal times.

With the pandemic, policy mattered but service delivery took on new importance. Overnight, one of the first things we had to figure out when we shut down schools in San José was how do we still deliver meals to the kids who qualify for free and reduced-fee lunches. The U.S. school infrastructure, like many other countries, is a critical service. Similarly, meal service delivery is critical to ensuring the right nutrition for children. So in a way, the pandemic framed my role in public service. 

That said, I’m very grateful for this opportunity. I think the hypothesis behind my fellowship was realised during the pandemic. Prior to business school, I had only worked in the private sector with growth-stage technology companies, such as Dropbox, and in very rigorous analytical environments, such as Bridgewater, one of the largest hedge funds in the world. 

They have very different types of culture, but in a sense, these experiences were helpful because they allowed me to work in environments where data was leveraged to drive performance. When I stepped in as Chief Data Officer for the City of San José, it was at a time when performance was demanded, and our mandate was to serve everyone equitably. This created many opportunities for data to be used by the City in ways that it hadn’t been used before.‍

The City of San José is one of the most diverse cities in the U.S. How does the City government make use of data and analytics to build more equitable public services?

San José is one of the wealthiest communities in the U.S. — we are the capital of Silicon Valley — and it is also one of the most ethnically diverse. 

I like to say that immigration and innovation go hand in hand. The diversity we have is not random. In San José, we have one of the largest Afghan refugee populations and we’re getting ready to welcome more refugees from there. We also have a huge Hispanic population and Vietnamese population. English is not the primary language in 50% of the households in San José. When I work for a city that really celebrates its diversity, I’m not serving a homogenous population, which means that access to resources differ, depending on the unit and the communities that our constituents are from. 

There are also differences in terms of their immigration pathway to the U.S.. There is a difference between someone from outside of the U.S., who got a full scholarship to study at Santa Clara University, versus someone who came to the U.S. as a refugee. Their ability to access the best that the City has to offer is fundamentally different. So we need to think about equity beyond just translating resources into different languages. That’s important, but we have to go beyond that. 

What’s really helpful about data and analytics is that they are a tool to span that gap. For example, governments love surveys. One of the things that governments like to do is to take a survey and translate it into 20 languages. And we’ll knock on people’s doors and ask them to fill it out. I’m a first-generation American — I grew up in the U.S. as a minority — and my parents have never filled out a government survey. I just want to be realistic with you — surveys do not give you the full answer. Honestly, it’s not just the U.S.. Many governments randomly mail surveys and hope some of them will get completed and returned. 

I find that it is more helpful to apply analytics to existing government databases: it tells you who you’ve already served instead of putting the onus on your constituents to tell you what they want. I think it’s actually more effective for governments to at least start with understanding who we were able to help for the last 10 years, understand where the gaps are, and then use that as a baseline for outreach. ‍

Can you elaborate further on what data equity is, and why it is more important today than ever before?

Data equity is using our constituents’ data ethically, and in ways that drive equitable outcomes for our residents — this is really critical. 

San José is one of the first cities to have an Office of Racial Equity. We formed this office in the summer of 2020, at the height of the Black Lives Matters protests. 2020 was a year of social justice reckoning for the country. And personally, I think it made us realise that the term equity can sometimes be a buzzword — you hear it used a lot by politicians and community organisations. But at the end of the day, we need to be rigorous about equity. 

I come from a very rigorous high-performance environment. I think I’ve always treated the term equity the same way that companies would be focused on revenue. In the private sector, revenue isn’t an ambiguous term. We know how much money is coming in and how much money we are spending. It is the same concept for me when we try to define equity for our different City departments — there needs to be consistency. And once we define it, we actually need to measure our performance against it. So that’s what has always been our approach. 

I developed a framework — a Data Equity Framework — that basically recognises that every City department is going to define equity differently. For example, the Parks and Recreation department is going to define equity differently from the Transportation department. This is also due to how different individuals and communities rely on these departments differently. We have to acknowledge that. My team developed the framework. We hired user researchers and service designers to go in and facilitate workshops to help each department define an equity objective. These are design thinking workshops with folks at every single level of the organisation. We define an equity objective and help them understand who they’re trying to serve. 

What sets our approach apart is that we actually use historical data, to essentially do a baseline assessment to understand how well we have served that community historically. And then, we commit to a roadmap so we know when we’ve missed a target, and what are things we can commit to moving forward to meet our objectives. I don’t think that this and tying equity to performance have ever been done in the City, so this is quite an innovation.

Data Equity Framework

Equity Objective

What is an equitable outcome for your department and programmes?

Equity Metric

How will we measure this? (Derived from our objective)

Monitoring & Evaluation

How do we ensure this long term? (Supported by City Manager’s Office)

‍‍Can you walk us through an example of how the Data Equity Framework is used within a City department to derive equity goals and objectives, and performance targets?

Let’s use the parks department as an example because every city has their parks and every city maintains public spaces. In San José, our Parks and Recreation department also manages community centres. In our city, community centres are places where kids can go after school and it’s where you have recreation programmes for senior citizens.

These are important places for a lot of low-income families. They are an important source of childcare — at least in the U.S. — as community centres are public spaces where kids can go to, and have really safe places to be. This makes these spaces especially important for lower-income communities. 

One of the largest youth programmes in San José that is run by the Parks and Recreation department, the Citywide Scholarship, subsidises the cost of swimming classes, chess camps, computer sciences, and after-school camps for low-income families. We had a hypothesis that we could expand it to serve more people. Using data to help us figure out how to do that, we found that wealthier parts of the city had nicer parks, community centres, etc — and so, equity became a matter of customising services for each community rather than delivering the same service to all communities.

For example, if a place had a tennis court and a pool, you would think that to keep things equal, we would have to build tennis courts and pools in places that did not have them. But when we engaged with the community by hosting roundtables for more than 300 community organisations, we found that the lower-income families did not need tennis courts. What they needed were community centres to stay open beyond 5pm because they could only pick their children up later in the evening. This was an example where we started with a goal, and using a combination of community engagement and data, we were able to learn exactly what was needed to achieve that goal.‍

‍

So once the challenges and gaps in achieving an equitable outcome are determined, how does your team and the City bridge the identified gap? How does this get cascaded down to the frontline City employees?

The thing about change is that it takes time. With frontline employees, it actually helps a lot when data is communicated to them. They’ve never had data. Frontline employees usually already know that something could be better because they’ve been serving communities. They know when the money runs out. They’re the ones who have to tell the families: “Hey, I’m sorry, you have to apply next year.” What was really helpful for our team was that when we communicate with frontline employees, the problems that we see from the data suddenly become more concrete, which is a good thing. 

We also noticed that 96% of people receiving scholarships live within a mile from the community centres, which shows that the government may not have been fully marketing its resources and that it hasn't targeted its marketing efforts properly. Sometimes you just walk by a building and maybe you see a flyer, right? So imagine if you take some of the private sector marketing principles, and you bring it into government, everyone could have everything. 

Having these kinds of insights — where you realise that we might accidentally be preferencing people who happen to live close to the park — is very important. If you look at housing prices in San José, the homes closer to amenities, like parks, are actually relatively better off, even if you’re in a lower-income community. It’s still the wealthier lower-income people accessing resources more than the poorer lower-income people. The data and insights made us stop talking about lower-income households and poverty as though they’re homogenous, which is not true.  

‍

City-wide Scholarship Recipients (2009-2019)

‍‍What would you say is your biggest challenge at work currently?

So, much of my work centers around influence. It’s about working very closely with the City Manager’s office — the counterpart to the Mayor’s office — which is focused on operations. Change doesn't happen overnight — we need to gain their trust and buy-in through influence and tangible outcomes. We've been able to partner with as many City departments as we have and build a reputation around our work because we put in the time to really understand the needs of day-to-day City operators. 

One of the biggest testaments we’ve had is that my team’s work has more demand than we can meet. As of today, we’ve already transformed six departments, and we now have community organisations reaching out to my team wanting to collaborate. I think it was helpful that our city’s governance approach is not top-down and that data equity came through soft influence. I started working on one project, and other people heard about that project and started coming to me and asking how they can understand their programmes better. I think that is a much more powerful change management approach than one that is achieved through executive authority. 

‍‍Some say that data is the new gold, and data-driven policies have demonstrated success. However, getting access to clean and useful data might be difficult. Can you share how the City of San José started collecting, organising, and making sense of public data in a way that is practical and sustainable? 

I don’t think that government data in its raw form will ever be clean. And I think a lot of best principles around data management and data collection have existed for no more than five to 10 years. In San José, the only way we’ve been able to scale capacity for that is by hiring talented data scientists. I lead a 20-member team, and we have about 10 to 12 data scientists working for us. 

The way we were able to scale that technical talent is by partnering with universities in the Bay Area, such as working with data scientists from UC Berkeley and San José State University. Our strong technical talent use Python to pre-process the data before analysis. I think this is a good way to scale capacity through talent. Imagine if I bought a very expensive software tool that did all of the data work for me. First of all, that software doesn’t even exist. Even if it did, that wouldn’t be accessible to a city with less resources. However, every city has access to technical talent since every city has access to a local university with students interested in this kind of work looking for practice. I think other cities should build similar internship and fellowship programmes with their local universities, hire data scientists to come into the city to help with pre-processing of data, and create something really exciting. ‍

There are so many different types of data being collected by government today. How does your team engage with different government agencies in San José given that they all have different needs and possess different kinds of data? 

It’s tough and this is something we get asked a lot. We do have a blog, and we are starting to write more about our procedures and we try to share them in an open source manner. The nice thing is that the way we manage the parks department data in San José is applicable to other cities because they all use the same software to complete and capture the data. There are many similar principles in terms of managing these programmes. The reality is that data is just a reflection of day-to-day operations so it’s never going to be that standardised. The things that make a data set interesting may be due to policy changes captured from six years ago. People usually think that our job is very technical, but you also need to have a genuine interest in the issue and be in the city to really make sense of the data sets. The example I gave you with scholarships was 10 years’ worth of data. Imagine how much a single city programme can change in 10 years? Can you imagine all the different City Council decisions made? I’m pretty sure that data set captured at least four to five different policy changes and different ways of funding. 

I think the short answer is that this process cannot really be automated. Maybe that’s actually a good thing. This is not a situation where my team is writing machine-learning algorithms to learn these data sets. If anything, we’re actually applying powerful summary statistics, and using that to tell a story and to shape policy. But the powerful summary statistics have to start with empathy: you have to start by reading the policy and understanding the history of the programme. ‍

Finally, what advice would you have for Chief Data Officers in other governments who have just started in their jobs? Are there areas they should look out for, or pitfalls that they should avoid?

What’s interesting about the Chief Data Officer job is that, often, we have to define our own role. I like to think about it in terms of what does the Chief Data Officer do? The Chief Data Officer is different from the Chief Innovation Officer. San José has had a Chief Innovation Officer for at least five years. We’re already on our second Chief Innovation Officer. The job of the Chief Innovation Officer is about bringing new things from outside into the city, but for the Chief Data Officer, it’s less about newness and more about working with what already exists, and finding new ways to look at what already exists. 

Cities need to have both functions as they’re fundamentally different. We need innovation, but we also need empathy and rigour and understanding of what already exists. In the examples I’ve highlighted, I haven’t spent anything  beyond paying my team — I don’t really purchase technology, that’s not my job. My job is change management: trying to shift culture and trying to use data to drive equitable outcomes. That’s a really unique and important mandate and I think all cities should start thinking about building that capacity.