The computerized crystal ball of crime
IT’S a joke among cynics that some police know a crime would be committed well before it happens. They even know who’s behind it just by looking in the mirror.
Jesting aside, there are new technologies that enable law enforcers to forecast crime. “Predictive policing” is gaining a significant boost from the rise in computing power and improvements in data collection, borrowing techniques from such successful forecasters of consumer behavior as Amazon and Wal-mart.
From crime hot spots mapping, data mining, geospatial prediction, to the more cuttingedge technology of measuring behavioral and physiological signals to detect the intent to cause harm or commit an offense, today’s police are better equipped to anticipate when and where the next crime will likely occur, allowing them to prevent and respond to future crime more effectively.
Since the dawn of the era of professional policing in the ‘60s, to the community- policing model of the ‘ 90s, and the emergence of intelligence-led policing after the 9/11 terrorist attacks, law enforcers have always had access to a huge amount of information.
So said Charlie Beck, Chief of Detectives at the Los Angeles Police Department, and Colleen Mccue, President and CEO of applied predictive analytics firm MC2 Solutions, in the 2009 article, “Predictive Policing: What Can We Learn from Wal-mart and Amazon about Fighting Crime in a Recession,” published in The Police Chief, the official publication of the International Association of Chiefs of Police.
The challenge is analyzing and [Guest article by Ric Saludo’s league Tanya L. Mariano]
col- leveraging relevant data to stay a step ahead of lawbreakers. By taking a cue from businesses using data analytics to anticipate consumer behavior and market trends, the authors believe that law enforcement can be less reactive and more pro-active, and can more efficiently allocate resources, especially during trying economic times.
It all started in 1994, when the New York Police Department launched Compstat, a method by which crime statistics are “collected, computerized, mapped, and disseminated quickly. Officers are held responsible for the crime in their areas, and all crimes, including the ‘quality of life’ infractions like loitering or public intoxication, are pursued aggressively,” according to a 1999 article in Government Technology Magazine, which features an interview with Compstat creator Jack Maple.
The technology was relatively rudimentary, but according to the article, “Compstat: From Humble Beginnings,” published in Baseline Magazine in 2002, it led to a 65.7% decrease in crime over a nine-year period.
Inspired by the New York experience, police departments across the U. S. started using Compstat as well, with some even building on its success to create their own predictive policing models. Memphis is often cited as one of the first cities in the U.S. to have successfully implemented a modern data-driven law enforcement style, and has become quite the benchmark for other cities follow.
Another example of predictive policing at work that was brought up during the First Predictive Policing Symposium in November 2009, as reported by the National Institute of Justice (NIJ), is the Police Department of Arlington, Texas, having developed a formula to help identify characteristics of “fragile neighborhoods.”
By identifying hotspots using data on home burglaries and comparing them with statistics on areas with housing code violations, the Arlington police found that areas with more home deterioration could expect more burglaries. Armed with this information, officers can more efficiently deploy patrols, and the neighborhood has added incentive to improve their environment.
Another program called DataDriven Approaches to Crime and Traffic Safety (DDACTS) uses geomapping to pinpoint areas with high incidences of crime and car crashes. According to the NIJ website, “drawing on the deterrent of highly visible traffic enforcement and the knowledge that crimes often involve the use of motor vehicles, the goal of DDACTS is to reduce the incidence of crime, crashes, and traffic violations across the country.”
One of the most controversial crime forecasting projects in development is the U. S. Department of Homeland Security’s (DHS) Future Attribute Screening Technology, one of the DHS’S Human Factors / Behavioral Sciences Projects. FAST, says the DHS website, it’s a “prototype screening facility that will measure both physiological and behavioral signals to make probabilistic assess- ments of mal- intent based on sensor outputs and advanced fusion algorithms and measure indicators using culturally neutral and non-invasive sensors.”
Translation: FAST analyzes bodily and psychological signs to assess if a person may have ill intent. The system uses concealed sensors to measure subtle changes such as heart rate and gaze. So far, tests on 2,000 subjects show that FAST can correctly determined malintent 78% of the time.
FAST is intended to be deployed in U.S. airports to supplement the current Screening of Passengers by Observational Techniques ( SPOT) program, which consists of DHS officers roaming terminals to scan for suspicious behavior. However, the public won’t be subjected to it anytime soon.
Last December the DHS document, “Civil Rights/civil Liberties Impact Assessment,” says “FAST is still in the research phase; it is not currently, and may never become, operational.” But the report allows: “Pending further testing and with appropriate operational limits, notice, and monitoring, we conclude that FAST could be deployed consistent with constitutional requirements.”
As always, keeping people a bit safer means prying a bit more into their lives. ( Excerpt from The CENSEI Report strategic research on crime forecasting technology. For a free copy of the Report with data, video and online research, email report@ censeisolutions. com.) Tanya Mariano is writer- analyst for the Center for Strategy, Enterprise & Intelligence, which publishes The CENSEI Report, strategic research on national, business and global issues.