Of all the ways technology is shaping law enforcement and public safety, perhaps the most cinematic is the ability to stop crime before it ever occurs. A range of applications, software programs, and services designed to see around the corners of the everyday crime pattern can help law enforcement plan more deliberately and channel resources accordingly. Although some examples may sound like science fiction, predictive policing regularly delivers real-world results.

“Predictive policing” can mean several things. In some cases, it centers on a mathematical equation designed to identify criminal patterns and trends in a given area. However, it can also be simply a way of making better use of data that already exist, sharing that information among public safety professionals, transmitting key items to officers in the field (and their mobile devices), deploying internal resources more efficiently, and leveraging the Internet and social media to play a greater role in solving crimes and understanding what’s coming.

Much of the time, the data needed for predictive policing already exist. In the case of a computer-aided dispatch (CAD) system, for example, meaningful data are gathered effectively, but can be difficult to access and organize into a pattern. Systems like CADs were designed to perform a task, not collate the information gathered as a result of that performance. Moreover, these systems are very often incompatible with systems from other departments or manufacturers.

Overcoming those challenges of traditional data collection systems is the key work that many predictive policing services undertake—breaking down those technological silos and mining difficult-to-tap but highly valuable veins of information.

“It can be hard to get answers from data,” said Nick Coult, vice president of product management for Lumen, a law enforcement search and analysis system developed by Colorado technology firm Numerica. “A lot of the time, it’s all stored in five different places in five different ways.”¹

One of the leaders in the field of predictive policing is PredPol, based in Santa Cruz, California. The computer predicting model, based on a proprietary algorithm, shows officers where crime is statistically most likely to happen next. It’s almost like weather forecasting for crime.

The model each day identifies geographic areas (as small as 500 feet by 500 feet) most likely to experience a pre-customized incident or group of incidents. The software then creates red boxes around those areas on a digital map, which can be distributed to officers and others for action and decision making. Though PredPol and services like it cannot be used as the sole basis for reasonable suspicion, it can help agencies determine where and when to put boots on the ground.

It appears to work, too. Within six months of using PredPol in its foothill area in 2011, Los Angeles, California, public safety leaders reported a 12 percent reduction in property crimes compared with the previous year, according to PredPol statistics.²

Lumen, also considered a leader in the field, works differently, attempting to connect the disparate ends of law enforcement data collection into one easy-to-use portal. The application (app) is capable of searching millions of records from a wide variety of databases, including CADs, record management systems (RMSs), and basically “anything that sits on a computer server,” Coult said.³ Type in a name or phone number, and results from a range of sources return in seconds. Users can share their findings, and the tool can visually map search results to make it easier to spot trends, which, in turn, enables more accurate prediction capabilities. Access to new data sources helps agencies cast a wider net—and work more closely with each other—when attempting to explore and address problematic areas.

“It removes barriers between analysts and investigators, and between ideas and the data,” Coult said. “It’s a huge accelerant… A customer knew they had a problem with a particular bar. Once they knew that, they could quickly ascertain the different incidents like DUIs, fights, and, in short order, they got a comprehensive list. As a result, they were able to take action and do things differently. And the incidents dropped off rapidly.”⁴

Another mobile app helps public safety professionals see neighborhoods in a new dimension as they walk the beat. Users of PublicEye, created by New Hampshire–based app development company Zco, can use the GPS feature on their iPads, smartphones, toughbooks, and other devices to see 9-1-1 calls, building pre-plans, and even geographically tagged tweets for a given area. For public safety companies, much of the information available on PublicEye eliminates the need for bulky paper files on buildings and similar records.

“Businesses already use phones and tablets, and it means big cost-effectiveness,” said Eric Sun, a PublicEye product specialist with Zco. “Our users are saving money. A lot of the time, instead of using an MPD, they’re using a phone or tablet—and they have better access to information. Officers outside the cruiser can access, and it’s geographically presented to them rather than sent as a text.”⁵

All of the information collated allows law enforcement to better anticipate potential issues. The app also has a streaming video surveillance capability, which the Brookline, Massachusetts, Police Department used as part of its heightened vigilance efforts during the 2014 Boston Marathon.

Predictive policing means not just capturing data more easily, but converting it into action. At IBM, the New York–based technology and consulting firm, SPSS Predictive Analytics crime prediction and prevention software helps agencies gather data from sources like CAD and RMS, as well as “unstructured” data from the web, such as weather forecasts and upcoming sporting events. This combination offers law enforcement a full picture of what lies ahead.

“It can take all of this information and make sense of it,” said Randy Messina, IBM’s government solutions manager for the predictive analytics software. “The text analytics and the structured data … help not just capture information but act on it.”⁶

SPSS then goes a step further with tools that help agencies decide how and where to allocate resources and devise new enforcement strategies based on identified trends.

All the tools, Messina said, can be tailored to a client’s specifications. Collaborating with community partners on the front end can help alleviate some of the primary concerns over tools like predictive policing, namely privacy and legality. “You can build your own model, and you can do it with transparency,” Messina said. “You can build your model not just with the input of officers, but from advocacy groups, judges, and prosecutors. You decide on different factors together. That buy-in ahead of time can be very valuable.”⁷

Several jurisdictions have realized big gains from their use of SPSS Predictive Analytics. The city of Memphis, Tennessee, saw a 28 percent reduction in serious crime and achieved an 863 percent return on their investment. Lancaster, Pennsylvania, reduced its crime by 42 percent over four years while saving an estimated $134 million.⁸

Predictive policing can certainly help law enforcement save money, in addition to mitigating crime. After all, having a better understanding of future trends allows for more efficient use of resources. This kind of analysis works in identifying internal trends as well as external ones.

Corona Solutions, a Colorado software company that caters entirely to the law enforcement community, offers Ops Force, a platform that helps agencies make decisions on hiring and shifts. Specifically, Ops Force works as a CAD search engine with a specific focus on staffing needs and trends, then helps agencies act on the information and create new rosters and schedules as needed.

“It tells them how many people they need and predicts staffing needs,” said Kimberly Roark, Corona’s vice president of operations. “If they feel they need more or fewer officers, this makes that decision-making process more than a gut feeling. It is based on real activity levels.”⁹

Ops Force also can help make the case to the public—and to those controlling the proverbial purse strings. “If the community wants, for example, a three-minute response time, you can tell them exactly what you need to meet that in terms of resources for staffing,” Roark said. “It can show exact response times. [For] agencies sitting down to think about how to deploy their people, this shows them how they can meet the demand for services more efficiently.”¹⁰

Other predictive technologies are also more specifically tailored. A service known as CROMS, created by Accident Support Services International (ASSI) in Canada, helps agencies better predict traffic patterns and collisions. “Our analytics portal allows for enhanced and real-time, ad hoc data analysis and search capabilities for authorized users, allowing for intelligence led policing to plan proactive road safety strategies for safer communities,” said ASSI President Steve Sanderson. “CROMS Analytics will query any field captured on the police report. Statistics are available in seconds. These records are a powerful means of tracking trouble locations in a community.”¹¹

As is apparent, the options for predictive policing technology vary widely, and agencies have to consider their needs and budgets when determining what products are the best choice. The right tools can help agencies manage budgets, effectively deploy resources, and even predict (and thus, prevent) some crimes. ♦

Notes:
¹Nick Coult (vice president of product management, Lumen), telephone interview, August 20, 2014.
²Donnie Fowler (director of business development, PredPol), email interview, August 18, 2014.
³Nick Coult, telephone interview, August 20, 2014.
⁴Ibid.
⁵Eric Sun (PublicEye product specialist, Zco), telephone interview, August 19, 2014.
⁶Randy Messina (government solutions manager, IBM), telephone interview, August 25, 2014.
⁷Ibid.
⁸Ibid.; Cindy Fay (public sector sales manager, IBM), telephone interview, August 25, 2014; Judy McCall (marketing manager, IBM), telephone interview, August 25, 2014.
⁹Kimberly Roark (vice president of operations, Corona Solutions), phone interview, August 20, 2014.
¹⁰Ibid.
¹¹Steve Sanderson (president, ASSI), email interview, August 25, 2014.

Please cite as:

Scott Harris, “Product Feature: Predictive Policing Helps Law Enforcement ‘See Around the Corners’,” The Police Chief 81 (October 2014): 44–45.