Predictive Software Helps Police Stop Crimes Before They Happen
Police are attempting to beat suspects to the scene of a crime with a computer program that indicates where criminal activity is most likely to happen.
It’s called predictive policing, and in much the same way that mathematical calculations can forecast earthquake aftershocks, this software uses past statistics and continually calibrated data to predict when and where crime will strike.
The idea began as a research project by an anthropologist, two mathematicians, and a criminologist at UCLA. After entering data about past criminal activity from the Los Angeles Police Department, the software they created was able to reasonably predict the location, time, and type of crime likely to happen in a certain window.
A software company was founded to develop and sell the program, called PredPol. The system is being used by police departments in Santa Cruz, Los Angeles, and Seattle, with a good amount of success so far.
“The feedback I’ve received is that there is appreciation that it has validated intuition or provided a new focus area that wasn’t known,” said Santa Cruz Police Department crime analyst Zach Friend.
New data is fed into PredPol each day. Based on that information, the software generates red boxes as small as five hundred square feet indicating where crime is likely to occur on a patrol map. The “hot spots” recalibrate with passing windows of time and as more data is entered. Officers keep tabs on the red boxes and patrol those areas more often than usual. The practice has led to more arrests, but departments believe it’s been even more effective at deterring crime.
The map and information are accessible from any tech device or on paper, and PredPol is run on a cloud-based software-as-a-service platform.
PredPol is mostly used to predict property crimes, car thefts, and break-ins, but Seattle is now becoming the first location to use PredPol to try to predict gun violence as well. The Los Angeles Police Department also is exploring including violent crimes in its program.
Andrew Guthrie Ferguson, a law professor at the University of the District of Columbia, has taken a special interest in the constitutional implications of PredPol. He says the departments using the software have told their officers not to rely on it for stopping a suspect—police still need “reasonable suspicion” to detain someone.
Though a computer program is free of human biases when it comes to identifying a questionable character, it remains to be seen whether the fact that a person was in a red box on a map will be a defendable explanation in court.
While predictive policing has proven to be a valuable and effective resource for police departments, it hasn’t replaced human experience or intuition—not yet, anyway.
