Socrata Interview Brett Goldstein City of Chicago

Brett Goldstein – CDO & CIO, City of Chicago

Interviewed on Feb. 5, 2013

Why the role of a chief data officer and why now?

I’ll answer it in two ways. One Chicago-specific and then one more broadly.

So why in Chicago? Why now?

Is May of 2011, we had Mayor Emanuel elected and one of the things that the Mayor campaigned on was open data and transparency. With Mayor Emanuel, it is far more than just talking about something on a campaign. It’s taking a promise and turning into action.

And, in order for something to be a priority and in order for something to be taken very seriously, there needs to be someone who can spearhead this piece. You have that foundational item coupled with Mayor Emanuel’s desire to have policy which is driven by empiricism, which is quantitatively supported.

So, you are starting to flesh out this role of a chief data officer and he felt that it was of such a priority that he made it a senior member of his office, as a political appointee in his administration. And, by doing that, he was able from day one to, first, have it as one of his 100-day goals but, two, have it be a critical part of what’s been the first couple years here.

Now, why is it important on the national stage?

Over the past decade, we’ve seen the role of technology change, at its most basic. In the early days, you had IT ops and your core role of a CIO.

You then saw the evolution to see things like a chief security officer. Why? Because cyber-security became critical to the enterprise, to the business, to administration.

But, what you’ve seen over the past few years is just how data can transform. Our relationship with data has changed in a variety of ways. Two of which, I think, are relevant to the discussion.

One, the advances in big data. We’ve seen how we’re going to move from BI to actual insights that can come from large data. But, in parallel, this idea of transparency through administrative data.

Our relationship with the public has evolved. Back a few years ago, your relationship was limited, with requests for information being answered in a pdf. Now the possibilities – and expectations – are very different.

You’re often dealing with the near real-time possibility of response coupled with the machine readability of the data. And when you have these sort of responsibilities, ranging from quantitative analytics to transparency, in a meaningful, honest way, you need someone to lead that way. I think with Mayor Emanuel showed a nationwide leadership move of making it a key piece of the administration.

There is the concept of the chief data officer in government but also in business. And, this is the person in the organization who is a senior leader and is thinking about data at a strategic level, not necessarily at an operational level. Is that something that you would agree with?

I think it’s a little limiting. The way I see a chief data officer is in three buckets.

The one that we know well, and is relevant to what you guys do, is transparency in open data. It’s to have an open data policy.

It’s another to effectively implement it and I think it is important for the chief data officer to have depth in this field to do it well.

Let’s look at how we’ve architected things in Chicago. Ninety-nine percent of the data that we put out self-updates. That is sustainable. That is an architectural design; starting to figure out how all these different schemas can be unified in a design.

Figuring out how to integrate in with somewhat legacy systems is another design piece where, in my mind, the chief data officer needs to be able to have that vision, but also a deep understanding of how to execute it correctly. That’s role one.

Number two is the role of performance management and KPIs. It is critical that governments have performance metrics. How do we do this?

You don’t want to create these sorts of standalone systems where everyone is constantly trying to generate stacks to put in there. What you want to do is create self-maintaining systems, which are relevant internally and externally so you’re accountable to the public.

In general, with anything thing to do with data and mathematics, you’re looking to follow the rule of parsimony, the rule of simplicity; to find simple, key metrics that help you evaluate success or failure to allow for intervention and improvement. How you build that platform to do it correctly?

Area three for the chief data officer is where we get a little more advanced. You guys know my passion is in the world of analytics, data mining, data modeling.

Someone has referred to this idea as the difference between data and information.

You can have a ton of data lying around but if you don’t build any information from it, it’s not useful. That is the other piece a chief data officer needs to push ahead.

The CDO needs to transform the agency from simply reacting to things they find in the data, to being able to proactively use the data, as well.

We have all the data, we’re now moving into the phase where we find the patterns underneath. We can make better strategic and tactical decisions based on all this information.

So, I have high standards for the CDO.

They need to be transparent. They need to be architecting these smart systems. They need to find ways to measure effectiveness but at the same time, we’re all still on the same path if we’re not leveraging the real power of that data and doing things that are really meaningful. And, when I say meaningful, I mean smart, empirical, quantitative action.

You mentioned these three buckets that make up the role of the CDO and one of the questions that we had was how is this different from the role of a chief information officer or a chief innovation officer? What are the differences between the roles?

I see the chief innovation officer as the person who has some smart, transformative ideas. They are able to be, potentially, a change agent. Rethink processes; bring some fresh ideas to improve systems.

A chief information officer, since I carry that title as well, is about IT ops. IT ops, for example in Chicago, is a massive undertaking and it has daily challenges. What does that mean? It means everything from email to ERP. The ERP alone is a massive system.

As we are re-architecting DoIT (Chicago’s Department of Innovation and Technology) in Chicago, we are pulling in analytics – the IT organization and use of data should not be separate.

A chief innovation officer has skills related to process and business development, but should be well grounded in the ability to have applied delivery.

The CDO is an interesting role where, I think at its heart, if you give it the skills of a data scientist, you’re getting at what you really need here. And, there’s a whole talk about, ‘What is a data scientist?’

A data scientist is a combination of someone who can, speak, write, and analyze, however with the requisite CS skills to be considering problems of machine learning, data analytics, stats, all of those pieces.

One of the programs I like that gets at what that could be is the Heinz Program at Carnegie Mellon, where they have a Ph.D. in machine learning within the public policy section.

By growing more of these folks who have those broader business and analyst capabilities but they understand the powers of the tools and data, we’re merging these worlds and we’re actually going to be able to do business a lot smarter. I think that is at the heart of the CDO.

You’re the chief data officer. You’re also the chief information officer and you report directly to the mayor. Is that correct?


Based on that description that you just gave of the CDO, do you see the data and analytics and the data scientists and the CDO shop, in general, having more influence on achieving policy in the future in government?

I suspect it will because we, for years, have talked about things like smarter cities and better urban planning. We are starting to actually leverage the power of these tools.

When you talk about big data in government, you’re starting to see the likes of Hadoop, MongoDB, and then people using tools like R to do things in a smarter way.

The year I was in the Mayor’s office, I sat in the policy shop. I was the quantitative representative in that group. When you’re offering the ability to do these analyses, it’s a game changer.

I think you’re starting to see people who are really motivated and want to come into government to make a difference.

Also, you’re seeing these open source tools. I would argue that the R Platform is as good as any for doing advanced analytics. If you give me R and Hadoop, and you hook into all these data sources, you can do incredible things.

There are some academic partnerships out there. The University of Chicago has just started this Urban Data Institute. These are people who are now focused on how to do smarter things with urban data. And, that’s a game-changer because instead of speculating about, ‘If I make this policy move, what will happen?’ you’ll be able to test this and model it. This is what we did in the private sector for the past decade.

Why won’t we use it those same powerful tools to try and do government better?

Day-to-day, what do you do both internally with different departments in the city, as well as your role in the community, the people you interact with outside the city?

I want to challenge people, I want to brainstorm with them, and I want to get some good ideas moving.

With that said, my agenda is twofold. One, I think government should be accessible. If people think they have good idea, it shouldn’t be hard to get in touch with me. I need to be out there, engage people, be responsive on Twitter.

Two, the Mayor has a vision of Chicago as a tech hub. We have enormous talent in Chicago. I think we need to support it, foster it, and be out there and involved. We’re doing some great things.

I have office hours at 1871, that’s one of our startup incubator, nest areas. I show up at meet-ups. Occasionally, I’ll talk at meet-ups. Or, be engaged with various community organizations.

We have great companies out here. I’m really proud of places like Clevercafe and GrubHub who are smaller companies doing smart things, coupled with companies like Morningstar. Morningstar is two minutes away. They are doing amazing things with data.

Additionally, within the community of CIOs and government leaders, I do my best to be in close contact with my peers. I want to hear people’s problems. There are a variety of leaders I talk to on a regular basis. I try and hear them and we try to focus solutions around that.

We need to get past the silos, think in an enterprise way, and then we’ll have effective scaled solutions. It is a lot of listening, understanding, de-conflicting, and at the same time, making sure that as my organization develops solutions that are enterprise solutions that come both from deductive and inductive thinking – because many of the challenges we have today can be a function of limited silos from the past.

What’s the profile of a future chief data officer in terms of qualities and qualifications? And what advice would you give those who aspire to it in their career?

Data science is at the core. Understand data. You can’t go out and talk about data, implement policies systems or strategy, without getting it. And, these are all learnable fields.

My undergrad is in government. After I got my B.A., I decided to get technology focused and it started out with me just reading a stack of books. People need to invest time in learning how things work.

It’s easy to run a tool that will produce an outcome. Understanding whether what you produce is valid is another piece.

When you’re talking to engineers or talking to people in community or getting out there, people respect depth of knowledge.

Working in government requires tenacity. It requires some strong executive sponsorship. Every day I wake up and I say, ‘I am going to do it right.’

The right way can often end up be the hard way. So, you need to have vision. You need be able to say, ‘I am going to choose the hard way’ and be motivated to follow through. The easy path is often not what’s best for community, for your residents, for your constituency.

Also understand that you don’t need to be a subject matter expert in every one of your departments but take the time to understand process, policy, the different agencies. None of us should have the hubris to think that just because we understand the technology we understand what this all means. ‘Good enough for government’ isn’t good enough. Our standards here should be the same standards we see in Silicon Valley. That should be our motivation because what higher duty do we have than our communities and our residents and doing things in a better, transparent way?

Are there courses being given? Are there resources out there to help people navigate the role of a CDO in a future career?

I don’t think anything is really out there on the role of a CDO. I think we’re going to see a lot in the data science space. If you go to iTunes University, they now have data science courseware.

There are multiple books, for instance, the O’Reilly books, or if you look at the Heinz School, there are some online presentations there that I had some of my fellows read.

There are copious tools out there. There are also the applied tools, many are open source, and there is great documentation.

The broader question, how to be a CDO? I think we’re still finding out. I’m eighteen months in and I’m trying to pave the way. It’s a good question.