Hendon Publishing - Article Archive Details
IBM predictive analytics help slash crime rates in Memphis
It sounds like something from a sci-fi movie: software that predicts where and when crimes are going to happen. But for the Memphis, Tenn. Police Department (MPD), the results of using IBM’s SPSS predictive analytics software couldn’t be more real. MPD announced recently that its latest law enforcement techniques using this crime-forecasting software has cut violent crimes in the city by nearly 29 percent since 2006.
MPD has been using IBM’s SPSS predictive analytics software for four years and is able to forecast trends, allocate resources, and identify “hot spots” to reduce crime rates with strategies such as directed patrol, targeted traffic enforcement, task forces, special operations, high-visibility patrol and targeted investigations. “Criminals leave a footprint,” said John F. Williams, the crime analysis unit manager with the Memphis Police Department. “There is a pattern to how they operate, and you have to detect that pattern.”
How it Began
Several years ago, leaders in Memphis observed a sharp spike in crime. The FBI’s 2005 report on crime within the United States showed that violent crimes on a national level increased 2.3 percent from 2004 to 2005, and in Memphis, that increase was 2.5 percent. Aware that traditional policing approaches and inefficient methods of paper-based crime analytics were becoming less effective, MPD turned to IBM’s latest technology.
MPD partnered with the University of Memphis’ Department of Criminology and Criminal Justice, which was already using the predictive analytics software, and developed Blue CRUSH™, or Criminal Reduction Utilizing Statistical History. The program is an innovative, evidence-based approach that zeroes in on particular areas based on crime trend data. Once focus areas are established, the agency utilizes all the resources it has available to saturate those areas to prevent crime and make arrests.
Analysts pull data each day using the IBM predictive analytics software, which crunches volumes of information showing incident reports in seconds, including incoming data sources from patrols, pertaining to type of criminal offense, time of day, day of week or various victim / offender characteristics. The agency is able to integrate IBM SPSS and a geographic information systems tool to both analyze and visualize data in the form of charts, geographical maps, and reports.
The initial pilot program with the Department of Criminology and Criminal Justice concentrated on the problem of gun-related crimes in a particular area of the city. “We knew we had an area of Memphis that was pretty much infested with gang-related activity,” Williams said. At the end of the pilot program, the data was showing that it was more successful than any zero-tolerance or saturation effort the department had ever tracked.
“We were very impressed,” Williams said. “We were able to pinpoint when we needed to place our officers in those areas. We saw a tremendous decrease in those areas. From that, we knew we had latched onto something successful.” Subsequent to the test period, MPD expanded the program over the entire city.
IBM’s Software in Action
The IBM predictive analytics software generates information on exactly where the crimes are occurring and why. By considering past criminal behavior patterns and then mapping their anticipated future occurrence, police departments are able to develop real-time intelligence, allowing for more effective use of forces.
Blue CRUSH uses IBM SPSS predictive analytics software to create multi-layer maps of crime hot spots and evaluate incident patterns throughout the city. Then the software connects the dots and establishes trends. The software enables Blue CRUSH to analyze an array of data across the city’s entire nine precincts or narrowed down to a single block.
Sometimes the software renders surprising results, Williams said. For example, he said the MPD might think that crime has been repressed in a certain area, but the software will lead them into that area before MPD actually recognizes that there is a problem. “The software has the ability to look long term where as we may be in a tactical mindset,” he said. “Sometimes the data may be pointing to specific crime that calls for creative ways of combatting it.”
Depending on the crime and the area, MPD may send unmarked units to catch criminals in the act, while other times, they deploy the Blue CRUSH vans to show that they have stepped up the presence in an area, which acts as a preventative measure.
However, sometimes because of the increased saturation and arrests in a certain neighborhood, the MPD will discover that the crimes have shifted to different areas—a concept known as displacement. If the MPD has been making a lot of arrests in an area and then sees a corresponding decrease in crime in that area, occasionally a neighboring jurisdiction will call them and say they have had an influx in those same crimes.
“That’s why we encourage other agencies to share records management,” Williams said. The more the neighboring agencies know about each other’s enforcement activity, the better they are able to anticipate a possible shift in crime.
Williams said each police station in Memphis is constantly assessing its focus areas and figuring in real-time analytics to combat displacement within the city. They scrutinize what they did last week, last month and the day before, and compare it to other statistics to determine what is effective and what is not effective. Commanding staff may have to change their tactics mid-stream as the goal is to be focusing on an area before the crimes occur.
One example of a highly successful Blue CRUSH campaign was in January 2010. Targeted police operations in Memphis’ Hollywood-Springdale neighborhood resulted in more than 50 arrests of drug dealers. Since 2006, carjackings were down nearly 75 percent citywide as of October, and business robberies were down about 67 percent. “It works,” Williams said. “The software is very effective in showing you where the crimes are. But then it is up to the user to deploy the resources to get the results.”
The residents of Memphis have had an overwhelmingly positive reaction to the Blue CRUSH program and the heightened police presence in their neighborhoods, Williams said. Blue CRUSH is also boosting police officers’ morale, as they are now making more “quality arrests” of people who are repeat offenders, Williams said.
“The officers are excited about what we’re doing,” Williams said. They are making positive arrests and seeing dramatic results, all while using about the same number of officers as a few years ago. “We are utilizing every available resource. We understand that crime has no boundaries. There’s no particular ward or precinct safe from being victimized by crime.”
MPD can now better guide daily decisions that address criminal activity and place officers in a better strategic position to respond to unfolding crimes. IBM’s comprehensive set of capabilities for data access, data preparation, analysis and reporting allow MPD to maximize the value returned from its analytical resources. Officers are able to take a more proactive role in the city’s ongoing commitment to crime prevention and public safety.
Blue CRUSH has greatly expanded over the years and now works in conjunction with MPD’s Real Time Crime Center (RTCC), a $3 million state-of-the-art crime monitoring and analysis venue that opened in June 2008. This new approach to crime fighting earned IBM and Memphis Police Department a 2010 Technology ROI Award from independent analyst firm Nucleus Research. MPD recorded an 863 percent ROI in just 2.7 months, an average annual benefit of $7,205,501, according to the study.
Key cost areas for the technology included software, personnel, hardware and training. The initial costs to implement the software were minimal for Memphis Police Department because it relied on expertise from the Department of Criminology and Criminal Justice. MPD does pay for ongoing software costs in the form of yearly license contract renewal and annual maintenance.
Hardware costs include servers and PCs needed to run the application. The greatest cost areas include personnel and the initial and ongoing training of civilians and officers.
The investment certainly appears to have paid off for MPD and the larger Memphis community. Williams said other agencies are constantly inquiring about the Blue CRUSH program and the IBM SPSS predictive analytics software at the heart of it. With proven results and an impressive ROI, MPD hopes other agencies will join them in the world of state-of-the-art policing using predictive analytics.
“We’re elated about how the software is working,” William said. “It’s been amazingly successful. We really think we have something that could benefit other agencies across this nation.”
Candy Phelps is a frequent contributor to Public Safety IT. She can be reached at firstname.lastname@example.org.
Photos courtesy of the Memphis Police Department.
Published in Public Safety IT, Nov/Dec 2010
Rating : Not Yet Rated
Click to enlarge images.