LPRC - Episode 27 – Using Risk Terrain Modeling to Focus Crime Reduction Ft. Dr. Joel Caplan (Rutgers University)

Episode Date: July 11, 2019

The post Episode 27 – Using Risk Terrain Modeling to Focus Crime Reduction Ft. Dr. Joel Caplan (Rutgers University) appeared first on Loss Prevention Research Council.

Hi, everyone. Welcome to Crime Science. In this podcast, we aim to explore the science of crime and the practical application of the science for loss prevention and asset protection practitioners, as well as other professionals. Co-host Dr. Reid Hayes of the Loss Prevention Research Council and Tom Meehan of ControlTech discuss a wide range of topics with industry experts, thought leaders, solution providers, and many more. On this episode, Dr. Joel Kaplan, Associate Professor at Rutgers University, discusses risk-trained modeling by examining spatial factors, describing the social landscape, and exploring the future of crime dynamics. We would like to thank Bosch for making this episode possible. Use Bosch Camera's onboard intelligent video analytics to quickly locate

important recorded incidents or events. Bosch's forensic search saves you time and money by searching through hours or days of video within minutes to find and collect video evidence. Learn more about intelligent video analytics from Bosch in zones 1 through 4 of LPRC's zones of influence by visiting Bosch online at BoschSecurity.com. Welcome, everybody, to another episode of Crime Science, the podcast. Today, as always, I'm joined by my co-host, Tom Meehan of Control Tech, longtime retail LPAP practitioner, a whiz in all things tech, as well as investigations. And then we're really honored and excited to be joined today by Dr. Joel Kaplan of Rutgers University. And Joel is one of the absolute global leaders, if you will, on environmental criminology, but particularly looking at how place and crime interacts,

operationalizing critical science that others like Dr. Weisberg have put out there and provided a ton of evidence for. But the Rutgers risk terrain modeling process is something I've long looked at as long as I've had access to it. There's a lot of good literature by Dr. Kaplan, by Joel and his colleagues in the works and, of course, out there in the literature. And so we're really excited to talk to you today, Joel, about some of the things that you're working on, how it came about, and most critically, how can a practitioner, a law enforcement agency, or a loss prevention department within a retail chain or other commercial place take advantage of what you guys are working on and how to think about

people and place interactions and what that means. So without further ado, Joel, if I might, welcome to Crime Science. Thank you so much for having me. It's a great honor. I look forward to chatting. Excellent. So I'll just dive right in and let's first take a look at Joel Kaplan, PhD. Why criminology, Joel? How did you get into it? And then when did you sort of morph, or was it immediately, into the environmental perspective within criminology? I got my undergraduate degree in law and justice. So early on I've always been interested in criminal justice. And prior to that I've worked as an EMT and so I've always been in emergency management right into the point in time where I worked as a police officer and a 911 dispatcher on

separate occasions. And when I started learning more and more about law and justice in the undergraduate program, I realized or solidified for me the reality that criminal justice covers and exists in so many life domains and affects so many different people through multiple degrees of separation. And I wanted to be part of that to affect the system in positive ways. And I, you know, tried to figure out where my place in that realm was. And I suppose this is where I ended up. And there's still a lot more to be explored. That's fantastic. And, you know, talking about environmental crime and spatial dynamics specifically, what kind of drew you to that? Who were some of the authors that you had read prior to that that kind of got you fired up, got you interested in sort of that specific area of criminology? specific area of criminology? Well, I find GIS and mapping exciting because it uses the other side of my brain. I've always been a logical thinker, and you know, in

academia, you do what you have to do in order to write and communicate and study effectively, but mapping and GIS let you use the artistic side of your brain. And it marriage, at least in my own head, where it gives me the ability to communicate in creative ways and also to ask questions in ways I wouldn't have otherwise thought about because GIS and spatial analysis display and communicate data and information in new ways and add new perspectives to those data sets. And it's from that new perspective that I and others are able to understand data and understand situational

contexts in new and innovative ways. So I suppose it's, you know, I think of GIS and mapping as an art form, and I've always enjoyed art. In fact, I remember when I was very young, I wanted to be a computer engineer. I wanted to work for Walt Disney, I remember thinking. And that was before it was even computer generated. And I suppose art has always been in my blood. And GIS and mapping give me the ability to be artistic and to use symbols and colors and communicate visually, as well as through text or numbers. Well, I've noticed, by the way, your kaleidoscope,

that it's an image, a powerful image that you've generated that is obviously very colorful, very artistic, but yet concise way to portray some of the types of places that might generate more risk, that might create different dynamics in a specific local environment. So would that be a good example? I mean, it's really a neat way to communicate, I think. Yeah.

In fact, the kaleidoscope was created actually by Les. He wrote a book probably now 30 years ago called The Crime Kaleidoscope. called the Crime Kaleidoscope. As Les and I began collaborating and when we began developing risk-trained modeling, we adopted some of his early work on risk and risk analysis and his image of the Crime Kaleidoscope, we adopted it to become the Crime Risk Kaleidoscope. And essentially what the Crime Kaleidoscope suggests is the cylinder of the kaleidoscope

represents the study setting in which you're interested in examining crime or other problems, threats to asset protection issues and so forth. And the parts or the pieces, the shards of glass within the kaleidoscope, features of the environment or environmental factors that come together in unique ways with every turn of that kaleidoscope. And as these features come together, they interact to create unique behavior settings for crime. So I was really taken by Les' analogy, and it's been a really good way to explain what risk-trained modeling does for different types of crimes in different settings, which is to diagnose

environmental conditions that lead to crime. Excellent. Tom, let me go over to you. Any introductory questions or comments? Yeah, actually, I think the kaleidoscope is awesome, and I encourage everybody, all the listeners to give it a look up. I had first been introduced to it and actually in your book and I have a lot of questions. I know the listeners from the retail world struggle with creating risk models. And I think in your book, when you talk about the theory of risky places, can you in lay's terms, explain kind of what that means and how a retailer could translate that into their world? Sure. So the theory of risky places, Les and I were thinking about the various papers and experiments and experiences that we've had with risk modeling applied by policing or as applied by police for issues of crime and oftentimes we come across the fact that crimes cluster which is well

documented in the literature and we've even written about the joint utility of risk train modeling with other techniques such as near-repeat analysis or hotspot mapping. And we recognize that new techniques don't have to replace old, but there can be different questions that specific techniques can answer, and some techniques that can answer questions better than others. And what we realized was that when we analyze crime clusters or the phenomenon of hotspots or crime hotspots, and then when that's used to anticipate future locations for crime, it assumes that crime doesn't move despite multiple engagements by police in these areas or in these hotspots if we assume that crime doesn't move that it's always

going to occur where it always has been then there's several issues that relate to you know sustainability how sustainable can police efforts be, and how do we understand what it means for crime to be occurring where it always has in the past. So we refer to that as exposures to crime, that is, previous exposures to crime influence future crime occurrence. And this could be caused or due to a number of different reasons, some of which could relate to contagion effects or the near-repeat phenomena where an instigator or an originator event influences similar subsequent behavior nearby within a certain

period and multiple incidences near repeats or repeat victimizations results in clustering which ultimately becomes hot spots so nearly hot spots kind of these stationary or the stable occurrence of crime at the same places over time is referred to as crime exposures. And vulnerability to crime is what risk-trained modeling helps to diagnose from these hotspots. That is, the underlying environmental conditions that attract illegal behaviors resulting in these crime outcomes. So a place can be vulnerable to crime because it has features in the environment that attract illegal behavior,

such as convenience stores, laundromats, gas stations, or the cliche dark alleyway, which would be defined as an alleyway with poor lighting. So when we diagnose these underlying conditions, it doesn't necessarily mean crime has to occur there, but based on the existing patterns of crime, these places are where crime more often not and we can use that diagnosis through risk modeling to identify places of high risk, which we refer to as vulnerable places. So the vulnerability exposure framework combines areas of high risk with places that have had recent past events, which we've found to be the most predictive model for locations where future is likely to occur. And that's the theory of risky

places. You know, example, Joel, I'm wondering is, as I sit here in the LPRC and UF research lab that we've got that replicates stores in different environments here, Security Operations Center, for example. And I'm looking out the window to my right, to my south, and we've got a one-of-a-kind here in Gainesville compounding pharmacy close by, very close by. And it generates a lot of traffic, a lot of interesting traffic,

if you will. Would that be a good example of how what's going on at our place and the risk or the threat that you might be exposed to here changes immediately because of that place and the proximity of that place? Yeah. So we've actually seen some examples where robbery or theft, having to take drugs. In some cities, we've seen pharmacies as significant risk factors for robbery because people steal or rob prescriptions and or the drugs after they filled the prescriptions. So if that occurs over and over again, then that's gonna be that area nearby,

a pharmacy is going to spot of crime. But if identify other pharmacies that may also be, have similar potential for becoming a hotspot, but not located in an area that's highly then not going to be as likely to become a hotspot or to be a location for displacement. So if you're focusing on a pharmacy that's known to be problematic, that has a high that's known to be problematic, that has a high concentration or clustering of these incidents that you're describing, then vulnerability could identify other locations that have similar qualities.

For example, maybe the pharmacies next to subway stops, near bars across the street or nearby, or some combination, which then connects back to our discussion about the crime cytoscopic interaction effects, which create these vulnerabilities. So you can have a vulnerable place that also is aggravated because of recent past exposures, but if you anticipate displacement or you try to address those exposures by suppressing crime, then understanding other vulnerable places can help you anticipate the likely displacement of these incidents and their emergence elsewhere. If you don't kind of address all of these issues at the same time, then any shift or any displacement could repeat and then become a new hotspot and then lead to new exposures.

Excellent. I appreciate it. That was kind of what we thought. We talk about it here as our members come through town and spend time with us discussing and looking at data and evidence and using that example by day or night that place or without that place interacting with other places there's also this sort of there's a there's a path that people walk between our buildings for some reason it's a natural a natural, in the Army we call it a natural line of drift. It's just sort of a natural pathway. And so to us that adds yet another variable that interacts with all these other things that are going on in place and time here

to create maybe a little more risk to an individual that might be at our place. Right. Fantastic. more risk to an individual that might be at our place. So, fantastic. Especially in the kind of loss prevention and commercial space, different types of retail or business infrastructure can influence behaviors resulting in crime differently. And in particular, you have specific companies that might be a type of facility. So you can have department stores or big-box stores or convenience stores or grocery stores that have different names and are owned by different people or different entities. But a grocery store or convenience store

can influence risks nearby and if you don't understand vulnerability, then you can simply shift those risks to another similar store by a different name. And I'm sure all of them would like to try to understand not just what's affecting this, but what's affecting behaviors resulting from the way in which our stores or connect to the situational context of behaviors nearby. It's about understanding the influence of people at places. And so, especially I would imagine in the business sector and as you're describing, this is not just a name brand, but it's understanding how the facility or the facility type connects

to the crime patterns that are occurring nearby so that they can all begin to work together to figure out ways to address these kind of comprehensively. That's fantastic. So I guess another thing, you mentioned LESS, and for our listeners, Joel, if you might, kind of who is LESS, and again, who are some others that you look to or that are excited and working on RTM with you? So Les Kennedy is a university professor at Rutgers School of Criminal Justice. He's the former dean of the School of Criminal Justice, and he's now been at the university for over 20 years. What some of your listeners might not know is that in my office on my wall, What some of your listeners might not know is that in my office on my wall, I have a diploma from the Rutgers School of Criminal Justice where I graduated from the master's program.

And it's signed by the dean, which is Les Kennedy. I was a student when Les was the dean. And the first GIS class I ever took was a crime mapping class taught by Les at Rutgers. And while I was there, I was working at the Police Institute. And Eric Pisa was a research assistant at the Police Institute. And he was the first person to teach me how to use ArcMap. At the time, it was version 3.1. That's excellent. Yes. Very familiar names to us. And again, for our listeners, we've worked with Joel Somm, always picking his brain on RTM and environmental criminology and the way we can use it as practitioners to really solve problems or at least address those issues more precisely.

And then Eric has also gotten on and involved with us on a couple projects as we've tried to stand them up. We just always, always want to make sure we are talking to the right people, the absolute experts on everything, as well as things that we might have expertise in to get a broader perspective, to get second and third opinions, if you will. Let me go back, Tom, to you. What questions do you have for Joel? So for a lot of our listeners who have different levels of expertise, some of our folks actually do have PhD level data scientists on there. If someone was really wanting to get into RTM, what's some platform or some software that they could start at a basic level? I know that it's a loaded question, but if you were saying to someone

that really wanted to start researching it, where should they start? Well, you know, to be honest with you, you have to start with the right mental state, to be perfectly honest. You have to be willing to ask questions about why, not just where. So, while risk-trained modeling is a spatial risk analysis and it's about mapping and geographic systems, it's really about trying to understand why and where. So, there's software applications that we've developed at Rutgers University to help automate the process of risk-trained modeling called RTMDX or risk-trained modeling diagnostic software which is available to anyone throughout the world it's a it's a online access on all platforms but risk-trained modeling was

originally designed as a free resource there remains free access to the software and all the resources to do risk-trained modeling as an analytical technique are still available on the website in our books, which describe the steps in detail for anyone who has the statistical or geographic information system, GIS skill sets, to do risk-trained modeling by hand. Great. And then I just had one other question, which is probably a little bit more technical. When I read one of your books, I remember reading about ACTION and kind of a guide for problem solving. Can you talk a little bit about

that? Because I felt like that was pretty easy to understand. So ACTION is another acronym that Les Kennedy came up with. It had a different meaning at one point but we adopted it and changed some of the what some of the words meant to fit into the way in which risk is used by practitioners in the field. And while this is used, often used by police, it is intended to be more of an agenda or a framework, a guideline for how to think spatially and how to apply the analytical tools like RTMDX to actionable information. So action is a focus on assessment, connections, tasks, interventions, outcomes, and notifications. And each one of these steps involves an interaction between the data, analytical outputs, human elements, which can be a variety of different stakeholders,

not just analysts, not just police, key stakeholders who can help to understand the situational context of crime problems and come up with solutions either through distant resources or innovative ways in which to address crime problems by forming what we call risk narratives. The key step within the action plan, risk-trained modeling is an analytical technique, and the results include tables and maps that diagnose environmental features that lead to crime. But what you do with that information is to form risk or stories that add situational context to the output. For example, in one city we focused

that we tried to study shootings and a risk-trained model identified convenience stores, laundromats, and vacant properties as significant risk factors. And where they all co-located in space were areas of highest risk. So we asked police, we asked members of the Mercantile Association, the local merchants, neighborhood groups, and various other key stakeholders within the community if this makes sense and if so why. And the consensus was that shootings were often drug related and the convenience stores were often open late, they're easy to

come and go, easy to loiter and in places where buyers are solicited by dealers to purchase. Then they're told to go to the nearby laundromats, which are usually coin-operated, not surveilled by a human operator, open late, not open 24 hours. So these are places where they can go to make the transaction out of sight. And then afterwards, the buyers can go to the nearby vacant properties, use the drugs, but also vacant properties are used as stash houses for drugs and weapons by the dealers. And when police and community members and the members of the Mercantile Association understood this risk narrative, it was much easier to build a consensus about what to do about this problem and how to intervene by focusing on these places and not just every single person that happens to

be located in these high-risk areas. So let me just say, that's a fantastic example, Joel. And maybe from there, I know you had another case study that you and I have talked about in the past that's coming up that involves using RTM to help the retailer and to help the local law enforcement partners of that retailer or retail chain much better understand what's going on at that place and why, and right around that place and the role that the nearby locations are playing in that. Is there something that you could, a way you could describe, okay, for our LPAP practitioners that are working with law enforcement, okay, how can I take all this great work by Joel, by Les, Eric, and others, and this particularly powerful RTM tool, how can I use this? I've got a handful of stores, let's say, in a particular market, and we're having different but serious problems there.

I need to be better at diagnosing these problems so I can be much better at treating them. Well, the first question a retailer would likely ask is, how does my store influence illegal behaviors, or does it? So whether you're a gas station, grocery store, or convenience store, or pawn shop, the question is, at the local level, is crime occurring around my store more often than it's occurring around other stores? And if I have multiple stores, even within the same jurisdiction, what's going on at the local level? What influence is my store having on nearby crime problems? So retailers that we've worked with and others that we've spoken to and those that we know about are thinking about the local context for crime, which frees them from the need to try to allocate limited resources for asset protection across using data that's often macro, such as the uniform crime reports or even citywide data if you

can get down to the municipal level. Because what we've found in risk-prime modeling, which is very consistent with theory and experience and evidence from environmental criminology over the past several decades, is that local environmental context matters, especially with different crimes and patterns thereof. So retailers have thought about how risk-trained modeling can be used for understanding the influence that their stores might have on nearby crimes. They've thought about how their stores might interact with other stores to mitigate or aggravate risk.

So they've considered the use of RTM for site suitability analysis when considering new store locations. And they've also considered using risk, and we've seen the use of risk-train modeling within stores. So there could be crime, many stores can experience or stores can experience a variety of different crime problems from motor vehicle theft to robbery to assault to overdose. All one site because they have parking lots and they have the store facility itself. And some grocery stores have pharmacies or liquor stores or even restaurants. So each one of these units or divisions within these stores can influence behaviors

differently and in some of these stores they exist all at once so the co-location is real and present. So understanding these dynamics has become very helpful for issues of asset protection and allocating human resources as well as technological resources from cameras and security patrols to different types of security devices within stores. And these are the kinds of things that allow retailers to think differently about the problems that they're having on their properties. just about what we know from the data we have. It's diagnosing the data we have to understand the problem and some of its underlying causes so we can respond appropriately. Another example that I really like about some of these responses we've seen is in terms of, you know, a lot of big retail chains have nonprofit, you know, where they give money to local community groups and organizations. And understanding the influence of your stores or other features in the environment that can aggravate risks of crime associated or connected or around your store allows these to prioritize giving

campaigns that focus on not just improving the wellness of the community which they often try to do but prioritizing giving so that they can be addressing the problems in the community that could also have benefits for the store in the vicinity nearby. So, for example, if vacant properties are aggravating the risk to nearby tailgates, then community giving might focus on encouraging community gardens on vacant lots. Yeah, I love that, Joel. And as you know, we use these five zones of influence and action as part of our operational model here. And we're talking today all about zone five, and you're providing ideas, thoughts, tool sets that could be used. Like you're saying, we're an

individual retailer or working with others or other commercial spaces and of course with the community including law enforcement you here you go you've identified crime generators or attractors that are approximate to you you're understanding the role they play now we want to give back to community but most most importantly want to enable and facilitate improvement in those areas so that there's more opportunity, non-crime opportunity and things like that. But by specifically focusing on, like you mentioned, an empty building and, okay, here's what we might do to mitigate that, this process, this community garden would not only enable and empower and engage local citizenry, working together, these people together, increase that efficacy, that collective efficacy, but also should help mitigate some of the risks, some of

the victimizers that may want to come to your place. They can't now stage. And I think real quickly, if I could, I had a note and I wanted to talk about that human and place, risky place interaction that you and I have talked about in the past. And as you know, we've had Dr. Tamara Herold on our podcast on crime science. We talked about risky place networks and how these places, what they are, how they're co-located, but how they enable and allow, facilitate would-be offenders, victimizers, to stage and to conduct their business, recruiting and all kinds of things that they do. What are your thoughts, if I could, Joel, on using some of the Risky Place Network concepts that she's developing along with what you all are working on?

developing along with what you all are working on? Yeah, well, we've seen, you know, some obvious connections are, you know, the virtual space of eBay or Craigslist, for example, where arrangements are made to make transactions, and oftentimes the convenient locations are picked. And sometimes it's these very parking lots of the businesses that we're talking about. And what you find is when you have certain retailers using sites, they choose it for reasons that are convenient for business, such as easy access to highway for deliveries and areas with large spaces for parking lots and areas where their patrons can get food or frequent to go to other places of interest. So when we think about problems occurring at specific retailers, it's not just that retailer, it's simply the location that's convenient that's providing an opportunity for these

incidents and the whole location of people at places. So therefore the interaction of people at places is exactly what risk-trained modeling is trying to diagnose in terms of environmental attractors and generators that bring people to these places to interact. But then when there's the networks, including social networks, that bring these people there, then what you realize is there's a lot going on. And trying to solve problems by focusing on just one piece of this very big puzzle is not going to be a comprehensive or sustainable solution. big puzzle is not going to be a comprehensive or sustainable solution. So if anything, risk and modeling and, you know, my study of environmental technology over the years has reminded me that there's always a lot of factors at play. You need to be willing and able to

understand them, not just technologically, but you need to think about the situational context and the various networks and pieces or factors that come together to create these opportunities. And risk-train modeling is one piece, you know, network analysis and, you know, various other aspects from criminology and psychology and economics and, you economics and various other fields of study also can play a big part in understanding not just where these problems are occurring, ways in which we can solve them. That's a fantastic, holistic view, Joel. And I know here we actually have some brand new signs we're working on because of that interface as you mentioned online selling meeting up at points in a brick-and-mortar retailers parking lot to conduct a transaction something you purchased or somebody you

met online or whatever those dynamics are coming from the cyber and the built social environments they're interacting they're taking place here things spill over from you a road rage incident into that parking lot and maybe even to that store. So understanding what things are there, what are the interactions, what's likely to occur. And the other huge point I think you made additionally on don't just think, though, one-dimensionally.

If I've got, and what we're all talking about here is, I think our listeners know, is if we've got a mosquito problem, we can just spray off or whatever brand you prefer, DEET on yourself and your loved ones. But you might at some point want to work with your community, with the other assets and resources to address where those vectors are coming from, where they nest and breed and come from. So we're reaching out now into Zone 5, and we're working collaboratively, but what else comes out of that wet spot? What else could be a threat in one way or another? But what else facilitates that?

Maybe there's trash cans that are upside down. They should be turned upside down or whatever. So I think you're right, looking at these things holistically, looking for efficiencies and opportunities to work together on multiple current or maybe upcoming issues. Tom, could I go back over to you? Any more questions or comments for Joel today? No, I think we covered just about all of it. I think one of the great things here that you talked about is it's not just simply about technology. So I appreciate the time and I enjoyed the book and look forward to speaking to you again and reading more. Thank you so much. You know, your mosquito example reminds me of a public health example where, you know, when there's mosquito problems, public health departments don't just go around and spray. They have advertising campaigns to encourage people to empty stilled

water and drain unused pools so that they can stop the... ensure that the sprays are one measure, but that there's a much more sustainable long-term plan. And I think that that very much relates to those, you know, what you described as the five zones of influence and action, in that stores need to think about how other stores and other places to affect them, and not only how they affect the community that they're in, because this is a very symbiotic relationship that every community has with its retailers and that the retailers have with all the different factors and features of its community. Excellent. Well said. And this is where, as you know, we're on that beachfront. We're right here

with the practitioners trying to make a real change and a difference, trying to safeguard vulnerable people in these places and spaces. And we've got to get it right and get better and better at it. And tools like your RTM risk terrain modeling concepts and software and interfaces are just fantastic. And your research and writing is helping inform us and add that in. So I just want to say thank you, Joel, very much for participating today. We always appreciate what you're trying to do and how you're trying to make a difference out there, and it is. So on behalf of my colleague, Tom Meehan, our producer, Kevin Tran, the 160 corporations that make up the LPRC community, and of course, here at the University of Florida, we want to thank you, Joel, for everything.

Thank you so much. I'm really inspired by your work as well. And I see it every day when I shop and explore the world around me. So it's a good team. Fantastic. Thanks so much, Joel. Thanks for listening to the Crime Science Podcast presented by the Loss Prevention Research Council and sponsored by Bosch Security. If you enjoyed today's episode, you can find more crime science episodes and valuable information at lpresearch.org. The content provided in the Crime Science Podcast is for informational purposes only and is not a substitute for legal, financial, or other advice. Views expressed by guests of the Crime Science Podcast are those of the authors and do not reflect the opinions or positions of the Loss Prevention Research Council.