Samir Saini, CIO, City of Atlanta

Samir Saini, CIO, City of Atlanta, says using the IoT (Internet of Things), big data, and analytics allows cities to forecast what will happen and to take action. He explains that while it is impossible to solve every problem within a city, offering the data to other organizations within communities allows for more improved collaboration and efficiency. He discusses the outcome of new undertakings such as the SmartATL and North Ave Pilot Project, as well as plans to power cities and to make infrastructure smarter for years to come.

To hear this interview on The Peggy Smedley Show in its entirety, log onto www.peggysmedleyshow.com, and select 10/31/17 from the archives.

Smedley:
Samir, I’d love for you to give us an overview of the SmartATL and what that really means for anyone who wants a real definition of the pillars and things you are doing in Atlanta?

Saini:
Sure. The thing about smart cities is it’s the buzzword of the year. It has a lot of different definitions to different cities. For us, we define smart city overall and SmartATL, which is just sort of our label of it, as meaning how we, or really any city, uses IoT (Internet of Things) and big-data analytics and machine learning, AI (artificial intelligence) to improve quality of life. It’s just about using the latest advances in ICT (information and communications technology) as well, IoT, and big data, to make life better in our communities. That’s sort of the core of it.

Smedley:
I think it’s interesting because in the city of Atlanta—you call it by descriptive and diagnostic and all these things—you use the information to really explain and to really forecast and recommend information for cities and businesses. Walk us through that, because I think that’s the power of what a smart city should really be about.

Saini:
Absolutely, so I’d summarize it in three ways. The first is using the full power of big-data analytics for the city itself to move from being reactive to proactive, which is really from an analytics sort of perspective, moving from descriptive analytics, which is simply what we’ve always done, quite honestly, which is able to answer the question, “What happened?” And really go up the chain from descriptive to predictive, which is being able to forecast what will happen, ask the question, “What will happen?” And to have an answer that’s reasonably accurate.

Then even further, quite honestly, it’s going to the prescriptive, which is have these analytics tools not only forecast what will happen, but give me a prescription, a solution, for addressing the forecast of what will happen, so it’s about really going up that evolution from descriptive to prescriptive for the city, which is huge, because then that means the city doesn’t have to just wait for a water main break or a pothole or crime to happen. We can predict it, and then we can respond in a way that’s significantly more proactive, which ultimately we believe improves quality of life.

Smedley:
Let’s talk about that further, because that’s an important point. It’s not only important for the citizens, but it’s important for the businesses and for economic success that you have overall, and the engagement that you have that everybody who’s living there and the well-being that comes together?

Saini:
This is where I think our smart-city strategy deviates a bit from where most cities are. Most cities are … Just call at smart city 2.0, which is what I described. They are deploying solutions that are pretty innovative, using IoT or not, and doing this level of analytics, moving from prescriptive to predictive, and are able to really be a proactive government. What we’re doing it differently, I think, it is what we’re now labeling smart city 3.0, which is first off sort of accepting the fact that we can’t solve all the problems of the city.

What we can do is offer all the data that we’re aggregating from the sensors we have, on servers, their in-house or from third-party sources, and make that available to the rest of the community, the business community, our college/university system, and the end citizens, so that they could use that data to help us with solving some of these problems, but also fueling economic growth as well, and ultimately, help us achieve the end goal, which is improving quality of life. It’s this notion of a city as a platform, and really establishing that platform and growing it and exposing and democratizing the data in it to these various groups.

Smedley:
Are you finding that you are engaging the citizens when you give them the information? Because you know, sometimes people don’t want information. What’s been the biggest success? That you’re giving everybody the data? You know, sometimes too much data is too much data, but now they’re seeing things so they can make better decisions both in business and personal ways of using it from a utility perspective. What’s that greatest success you’ve had so far from that?

Saini:
Sure, so I don’t think we’re there yet, and I think where most cities are with their smart-city implementations and progresses, initially focused on using the data to support the departments within the city itself to improve efficiency and service delivery, and overall effectiveness of governance. It’s the internal look first. But the next step is, of course, expose that data to citizens so that they can get value from it and tackle some of these problems at a local level. Some of the examples I have are really around how the city’s using this data to move from descriptive to predictive and improving various aspects of quality of life. I could share a few with you if you’d like.

Smedley:
Yeah, let’s do that, because I think that’s where people go from this idea where you talk about 2.0, 3.0, so the city manages itself, and that makes citizens happy. Once you make citizens happy, that’s the next level, right?

Saini:
Oh, I know. Of course, right? We’ll take an example of a data project we’re working on with Georgia Tech. First off, I mentioned, the power of having the platform is to provide access to it to other groups, not just the city itself. In this case, we give access to a whole bunch of data on this platform to Georgia Tech to help us, particularly with our police department, on basically crime correlation. The project requires some deep dives on big-data analytics and machine learning to basically look at all of the case reports we have in the police department, which are in large parts free texts, and have a machine learning algorithm using natural language processing, scan through it, extract features from it and actually, ultimately, calculate the probability of correlation between case reports.

This is a big deal because today it’s actually very difficult to correlate separate cases, and when crimes occur and a case is … An investigator’s assigned to a case, it’s difficult to know which investigators we should group together because we don’t know the correlation across so many cases, because of the sheer volume of number of case reports. Georgia Tech took on this task. They’ve developed this algorithm. It works. We’re actually going live with it very soon, and the notion here is for the first time we could fundamentally transform the way cases are assigned to investigators, find correlations across cases, increase solvability of cases, and what that ultimately means is preventing future crime. It’s an example of where we’ve extended the platform to Georgia Tech to do something we couldn’t do to help our citizens boost public safety in this case.

Smedley:
You’re kind of looking at patterns in a lot of these things, right? I mean, that’s where machine learning comes in, that most citizens don’t understand where the IoT comes in and big data and analytics and all of this is what you’re doing with these case reports. It’s seeing things that you might not have seen, and that’s what these case reports are bringing out.

Saini:
That’s exactly right. You know, one example we use is when we’re training the algorithm. There happened to be this string of 17 burglaries we had in one of our neighborhoods many years ago. We ended up closing the case the old-fashioned way. But we found out…We ran the algorithm and it detected the high correlation between these 17 cases, and when we got into trying to understand how, what it found was that the term silver was in each of these 17 case reports, but not just silver in terms of the color, but as a metal, a semi-precious metal.

The algorithm found the term silver in these reports and understood how it was used in the reports, and then spit out this high correlation between them. It turned out the burglars were after silver. They’re stealing silver in all these homes. It’s something honestly, humans can’t, we can’t do, that can make a huge difference. And we believe will make a massive difference in improving public safety in our city, and it’s about patterns at the end of the day, yeah.

Smedley:
Tell us briefly about the North Avenue Pilot Project.

Saini:
As part of smart cities we are learning and experimenting. What we’ve done is identify this corridor. This one corridor within the borders of midtown and downtown. It’s about five miles, 18 signalized intersections, and it’s really a living lab. We’ve invited a number of tech startups, big companies and small, to deploy their IoT technology on that street, push the data, make it accessible to everyone, and really start examining that street, that corridor, in ways we couldn’t before because of all the data we’re collecting. That’s one thing. It’s a living lab. It’s right in the backyard of Georgia Tech as well, so it’s sort of this outside living lab for Georgia Tech. We recently deployed an adaptive signal control system on it, so there’s a true AI signal control system on the street, which is really exciting.

Peggy Smedley, host of The Peggy Smedley Show