HxGN RADIOPodcast

Building a Data Culture in Public Safety Agencies

JW:  Hi! Thanks for tuning into Public Safety Now on HxGN Radio. I’m your host John Whitehead, vice president of sales for U.S. Public Safety here at Hexagon Safety and Infrastructure. There’s a huge explosion right now for big data in today’s public-safety agencies, and while there’s a lot of opportunity there, it’s also creating a lot of challenges. So today we got a great conversation that we’re going to have with Jack Williams. He’s our product manager here at Hexagon Safety and Infrastructure, and we’re going to discuss today how agencies can start building their data culture, transform some of their decision-making projects, and just kind of how does this big data fit into policies, their operations, and hopefully a whole lot more. So, Jack, welcome. We appreciate you being with us here, and I’m looking forward to our conversation.

Jack W:  Hey, John, great to talk to you, and I look forward to talking more about this topic area, which is very near and dear to my heart.

JW:  Yeah, I know it is. You and I have had several conversations before about this, so I’m always excited to talk to you about reporting and analytics. You know, it’s this business intelligence. I know that there’s some buzz and different wording all around the industry. I find it interesting how this thing started. You know, if you go back, you know, CompStat, in public safety and policing. That’s what we always heard, right, was CompStat reports this and CompStat reports that, and if you look back, that’s a mid-‘90s terminology where that thing started— But it was interesting. Essentially, it started off as just a bunch of pins on a map, just getting that information out there. It’s just grown so big and so bad that the data that’s out there and the ability and the computer systems to be able to use this data is amazing. So, tell me a little bit, Jack, from that, how does this term data culture, what does that mean to public safety from what we started in the ‘90s to where we’re at now?

Jack W:  Yeah, so, you hit on it, John. With the explosion of big data, it really provides agencies with a vast new resource that can help transform the agency, helping them to build smarter systems that drive efficiency and probably most importantly improve safety. And in order to do this—I like to use this term—agencies need to establish what I call a data culture. In a data culture, what it is, what does that mean, right? It encourages decision makers to focus on the information conveyed by the existing data and make decisions and changes according to those results. This means, what this really means is agencies have to let the data speak for itself and trust the steering of that analysis. So being successful at establishing a data culture at a public-safety agency requires the active participation of users traditionally not involved in “analysis.” So, in order—

JW:  So, it kind of gets it into the hands of everybody, is what I’m understanding. And it doesn’t have to be some guy sitting in the back office that’s got all of his brainiac type of degrees on the wall, right? I mean, it kind of gets this into the our hands, the common guy’s hands.

Jack W:  Exactly. That’s what—and the cornerstone of it all, John, is open access to analytics for any data, that’s the cornerstone of establishing a data culture. So, you need to have all that critical information in one place to answer your questions, and it needs to be ready for analysis. It can’t just be that raw data. That’s step one, and that’s the cornerstone of establishing a foundation for data culture.

JW:  So, here’s the thing, right? When we talk about this, I always think in the back of my head, it sounds easy. It sounds like, “Oh, data culture. I can just put some data on a sheet. I’ve got all this great data here, I might as well just throw it on a paper and just get it out to my command staff,” but there’s some challenges when it comes to this, isn’t there?

Jack W:  Yeah, I wish it was that easy, and we’d all be in a big old data culture right now. But yeah, there’s a lot of challenges associated with turning this vision into reality. It’s really difficult to take, even though we’re rich in data and our agencies are rich in data, they are often poor in information. That’s a famous line that you hear quite often. So, it’s difficult for some of these agencies to turn their data into real insights and answer real business questions. You mentioned CompStat earlier. CompStat is something that was introduced in New York in the ’90s and was revolutionary at the time, but it’s just, it’s statistics over time, and, you know, it’s listed out typically in sort of a table-based format and pins on a map associated with that. That’s good and that was step one in helping to answer real business questions, but even that by today’s standards is useful, but there’s better ways to view and visualize and actually get answers to your question, not just see the results. Inefficiency, traditional ways of doing analysis are very challenging because there’s typically not a repository of data that has, what I’ll say, analytics are reporting ready. Sometimes it takes a week or two weeks to get the data that an analyst might need to help make some decisions and change processes and be proactive. But a week or two weeks is way too long in relying on people like IT to query the database and try to make sense of it and then hand that off and then the analyst is using Excel spreadsheets, they’re using some other piece of software. I mean, an analyst seems like they have a toolbox of analytic software, John. It’s a hodgepodge. They’d have to go to so many areas to really do their analysis and use so many tools that it’s very challenging.

JW:  It is, yep.

Jack W:  There’s other things—

JW:  It seems like there’s—it’s still challenging, right? And we don’t want to make light of that. There is no easy button is what I’m hearing you saying, and it sounds like that there are definitely some tools out there, and I think we’re going to talk a little bit later here about machine learning and some AI, how we can kind of take that in. But before we get down that road here, you know, one of the things that I was always interested in was kind of hearing the end result. I know right out of the gate you mentioned New York City with their CompStat and when they were doing that. You know, you can do some quick searches online and see the results of those meetings. And even then, as they were doing these meetings that they were having, and these were mandatory meetings that they had to sit there with pushpins and maps and do these CompStat meetings, they started bringing crime down. They were seeing back in the ‘90s, with these manual modes, 60% crime reductions, and murders dropped down. And it’s amazing to me just some of the communication in utilizing this data, how it’s making a difference. So, those challenges we just talked about, we need to turn those into meaningful inside- and information-based decision making. So, tell me a little bit about that. How can I take this raw data, using some of those challenges and overcoming those challenges, but then when I get the report and I have it in hand, how do we turn that around to where it becomes meaningful?

Jack W: Yeah, so how do we take those challenges and make something meaningful so that we can make better decisions? And that’s really the key. You can have the best analysis in the world, John, but I tell you what, if you can’t convey those results to the end user in a digestible format, it’s just not going to work in an optimal fashion. So, a couple things. So, one thing is supporting your decisions with data. A lot of times, the output from reporting in analytics supports what the officers or first responders know and objectively understand. But having the data to back you up is the evidence you need to objectively propose changes, measure results, and measure the impact of those changes. So, having data back you up and support your decisions in an objective format is a benefit. Second is we can start to objectively assess new programs and policies. We can start new policies, objectively measure those, and continually refine and contribute to a proposed action. We can create shift and patrol plans that optimize our time and efficiency but that also consider officer well-being, which is a major topic area lately, a major area of concern. Validating and being transparent with our stakeholders, validating results to support things like public campaigns and funding requests, those are very big. Unlocking your data with easy access to all of your users, being able to have that data on demand via a web browser, letting people do self-serve analysis, all of these things are ways in which data analysis and analytics tools can turn these challenges into meaningful information to help make decisions.

JW:  Nice. You know, you keep using the word objectively, and I think that’s an important term. I don’t want to overlook that word, because in doing so, in having all of these tools, in having all of this data here, I think sometimes it’s easy, right? We’re just human nature to be able to say, “Oh, maybe this is a bad part of town. We need to have more officers patrol that.” Well, is that real? And how objective is that, right? There’s naturally going to be some biases, I would think, in some of these decision-making processes. But to your point, if you’re using this data and you’re viewing this data and you’re looking at it objectively, to me that’s a win all the way around. It says that these are areas that I may not have even thought about that I need to handle differently, patrol differently, maybe from the fire side I need to put more resources in that area. Again, the data doesn’t lie. The data is what it is, and I think that’s a key point. So talking about that from a personal bias point of view and keeping that out of there, I think that leads right into what we’re hearing about now and in some of this machine learning and AI, to be able to take this data and give us better information with it, but to do that through the machine itself. What’s your perspective on how those technologies are going to help us?

Jack W: Tons of potential, John. Tons. I view artificial intelligence and machine learning—now, here’s the way I view it. I view it as a way to help and advise public safety agencies, the users themselves, the first responders, not necessarily replace our first responders or dispatchers or call takers. So that’s first. I view it as a way to help and advise, not replace. Second, you know, if we just look at some of the bigger trends, what are some of our biggest challenges? Current call takers and dispatchers are overburdened. They retain—I am amazed. I am amazed when I go on site, John, and realize how much information call takers, dispatchers, first responders retain in their heads just to be able to leverage something like a machine learning to provide a second set of eyes so they don’t miss anything. I think a lot of times we forget these are human beings. Their stress levels are extremely high. They can have bad days. Folks in this industry are becoming harder to recruit, train, and retain. So, anything we can do to help them and that burden that’s put on their shoulders with a second set of eyes as an advisement, I think is one of the major areas where AI and machine learning can help. Second, would be more understanding the big trends. Looking at historical data, comparing that and correlating that with your live incoming data, sometimes, you know, you’re just focused on a particular agency, maybe fire. Maybe you’re just focused on the dispatching side or the call-taking side. Having the system itself leverage AI techniques and machine-learning techniques to help get a bigger view of the picture in trends, things like a major event might be unfolding, and the system can help detect that and just send up a notification to say, “Hey, we are noticing a trend that there could be something happening.” And sometimes—yeah, people do this today, but usually that’s a human analog type of thing to be able to do that. Enhancing situational awareness. We want to provide our first responders with as much information about the incident and event, et cetera, as possible, for their safety and for the better response to the actual citizens or folks involved.

JW:  Yeah. I like that, Jack, because what’s funny is that the more technology and the more tools that we bring into public safety, I think that there’s this thought that, “Oh, I’m going to give them another tool, I’m going to get more information, I’m going to have another technology, and let’s bring that into—in this case, we’ll just use the dispatch center.” I think you brought up a really good point here. There’s a lot of information going on through that dispatch center today. I’ve got a caller screaming in my ear because maybe my response wasn’t fast enough. I’ve got another caller that’s screaming in my ear because they’re having some serious issue in their life and they need help. And on top of it, I’ve got all this technology sitting in front of me and all these data feeds and information coming in. And the fact of the matter is I don’t want to bring machine learning in and AI in learning to take the place of people. I think sometimes that’s the first people think is, “Oh my god. You’re going to automate my job right out of existence.” No. Actually what I want to do is I want the machine learning to enhance that position. I want that dispatcher, that call taker, to be enhanced based on that machine learning and having that AI there, because to your point, and I’ll expand on it, I don’t have time to sit there and scroll through data feed, data feed, data feed to get that one little nugget that I need that’s going to help me respond or get my guys and team out that much quicker to help that citizen that’s on the call.

Jack W:  Exactly, John. You nailed it. And that’s the key. I think it’s to advise and to be assist. And, really, well-being, helping lower the stress levels, helping us to be more effective and efficient in our responses. We can use it to automate what I’ll call redundant tasks. But there’s still always going to be, I’m a firm believer, and definitely the human element being at the center of public safety and first response.

JW:  Yeah. That makes sense. So, okay, the next think that pops in my head—it’s something, as I’m out there, and you probably hear this a lot, but I love the movie, but unfortunately Tom Cruise made it tough on the future of public safety. So, I keep getting questions about minority report. So, let’s take that into some more reality.  Today you’re hearing that called predictive policing or predictive analytics. It’s trying to say, on this corner of this street at this time of day a mugging could occur. Someone could get carjacked. So, it’s this predictive analysis based on historical information. What’s your take on those?

Jack W:  Yeah, I’m going to give you my personal take on this, John. Whereas, what we were talking about earlier, which is taking artificial intelligence and machine learning and basically solving traditional problems in public safety using these new methods, predictive policing and that term, to me I view that a little different.  It’s a popular buzzword, in my opinion. There are some, I guess, concerns with it. One is anything that predicts more of a location, kind of like what you said, taking into the compactor of things, they call it risk-terrain modeling. Lights are out on a street or broken windows and this and that, that there’s a high likelihood that robberies will occur or vandalism will occur. Those are okay.  I’m okay with that, and I think there is some data-science principals and algorithms that could produce solid results.  But where I get a little iffy is when we’re using the term predictive policing to do what I’ll call some level of profiling, either individuals likely to commit crimes or individuals likely to be victims of crimes, that’s where—and coming from a data background, there’s a thought, and I tend to believe it, data inherently has some bias. And when you start to use data to profile individuals, you can end up reinforcing the bias inherent in the data. I don’t know if that makes sense to you, John. I think to do it, I think there’s some potential there. I think that, though, when it comes to predictive policing, it requires full transparency. It can’t be a black-box algorithm that’s making these decisions. You’ve got to be able to explain it. And you also got to be aware of the tendency for if there’s bias in the data, it will reveal itself in the algorithms itself. Personally, if it’s away from the individual and focusing less on the profiling aspect, I think it’s a promising area. I’m just a little personally hesitant when it comes to profiling people and likely to commit crimes and likely to be victims of crimes.

JW:  Yeah, so it sounds like it could be a slippery slope, but it’s definitely one that we got to kind of walk in with being cautious, not just taking it for granted that the data that’s coming out is gospel, that there is going to have to be some objectivity, I think, in looking at that and taking that forward.  So, it sounds like that’s still a work in progress that we’re going to just kind of slowly consider as we move into the future. In other words, no jet packs on our police officers [unclear], okay. All right. So, I know you’ve got some exciting news here. You were just out recently. I mean, you’re on the road a lot, man. You’re a busy guy as well, so I know you were just out, and you just were on site at a customer site, doing some training with them from an analytics point of view. So, tell me about their feedback. What kind of feedback were you getting, because this is fairly new for that customer. But also, how do you think that they’re going to use it in the future? Because it’s one thing to go in and train and just show everybody how it is and they’re all excited about their new reports, but how do you think they’re going to use it, and what was their feedback for it?

Jack W:  Yes, at first, I, you know, in my role, I was fortunate enough to actually get on site, be with the customer, participate in the training itself, and really see firsthand how once they understood what the platform could do, how it really opened their eyes. The feedback was excellent. They now have a single consolidated view that provides access to interactive reports and dashboards that they can share very easily and provide access to the folks who need that data. So, it enables—it’s that cornerstone piece. We put the cornerstone in the data culture by providing that data, a single consolidated view of our data. You know, they actually looked at me, and they said it’s so great to finally have on-demand information that they can rely on and trust—that’s a big key here, trust—and that they can share. And they wanted this, they thought they were going to get it, and in reality they now have it. So, everything—we had folks in this class, everything from just a standard qualified user to command staff to council members to chiefs. We’ve got those reports and that information that they need covered now. This is what they told us. All of our training participants were able to walk away with the reports they need, the analytics they need, and they have a foundation to—once they’re up and running, now they can just go with it, John. So it is, I mean, basically we instilled the foundation for data culture. It’s great. It makes my job rewarding when you can go on site and you can see the excitement in people’s eyes. I know I’m kind of a tech guy, and when people get excited about data and some technologies, it can seem a little goofy sometimes, but this is actually impacting their jobs and their ability to be efficient and optimize and share information with their stakeholders. And they were genuinely excited about all the possibilities, and that just made my day and it’s made my week.

JW:  Yeah. That’s got to be a good feeling when you’re out there and you’re showing technology that’s actually making a difference and it’s going to be part of their workload. That’s exciting. So, all right, we only got a couple minutes here left before we kind of wrap this stuff up.  So, do you have any advice for some agencies? I know there’s a lot of, hopefully there’s a lot of people listening to this and thinking, “Yeah, this sounds great, Jack, and boy, that customer’s doing well with their new product,” but do you have any advice on how agencies that may want to build or embrace this data culture, what they can do?

Jack W:  Yeah. These could apply to anybody, John, not just potentially folks that we deal with as our customers, but any public-safety agency. First off, it starts at the top, right? You got to buy in. So, it requires buy-in at the top level. That doesn’t mean the top level has to understand every little nook and cranny of data analytics, but it requires an investment and a commitment from the top to become an agency that creates a data culture, that is a data-driven agency. I think, too, the other thing is, this isn’t just reporting. My advice is to look at this as an opportunity to transform your agency. Having data to objectively back up what you’ve subjectively know is a great thing, and it helps you to objectively measure impact.  Require, or encourage, the active participation of users who are not traditionally involved in analysis. Incorporate more analysts into your staff. I’ll be honest, John. I go up to Canada quite a bit. They have teams of analysts that work for some of these agencies. And you can tell. They’re efficient. So, invest in the human resources to have more analysts. And invest in analytics platforms and technologies. And don’t forget the key to all this, though, that cornerstone, is for a data culture or reporting or analytics is the power of a good data warehouse, giving you that single consolidated view of your agency’s data. Those are good general starting points to help create a data culture within your agency that can have transforming and lasting impact.

JW:  Awesome. Jack, those are some great tips. Man, this has been an awesome conversation. I love talking to you about this. I can see the—well, actually, hear the excitement in your voice in this case, so it’s always good to talk to you about this. I want to give a big thank you to our guest today, Jack Williams. And for more information about today’s topic, please visit us at www.hexagonsafetyinfrastructure.com. And to hear additional episodes or learn more, visit us at hxgnspotlight.com. And thanks for tuning in.

Welcome to HxGN SPOTLIGHT!

Hexagon is excited to announce the launch of HxGN SPOTLIGHT, our new content platform that offers original Hexagon videos, articles, podcast episodes and more! Check out compelling stories from key opinion leaders, explore customer successes, and dive into trends that are shaping a diverse range of industries.