Predictive-Policing

Hagerstown Police Department
Intelligence-Led Policing
Using Predictive-Policing Technology to Prevent Crime


Overview
As part of the Hagerstown Police Department’s two major initiatives, Intelligence-Led Policing & Community Oriented Policing, the agency began, in 2014, using the latest crime-fighting technology to support our efforts to predict and prevent crime.

Predictive-Policing is provided by PredPol, the predictive policing company whose software combines past crime data, criminal behaviors, advanced mathematics, and computer learning using three key pieces of information – type of crime, place of crime and time of crime. The software then generates, for every shift, maps with a series of red boxes measuring 500 by 500 feet, representing the most likely locations where crime will occur. This gives HPD officers an opportunity to be in the area before a criminal strikes instead of just responding to crime after the fact.
For the officer on the street, spending a few minutes in each ‘predictive box’ every couple hours, has been proven to make a difference in jurisdictions that are already using predictive policing.

Predictive Policing
(Map showing the top three prediction boxes city-wide. Each officer receives an individual map with the top three prediction boxes for their beat as well. The crime data is built using the crime types listed. These five crime types are used because they occur more frequently than a homicide, for example, and can be more accurately predicted using predictive-policing software models.)

Measuring Results
We will measure the success of Predictive-Policing by the number of fewer crimes in each sector of the city rather than by more arrests. Predictive Policing is more about crime prevention than criminal apprehension. And, less crime equates to fewer crime victims.
 
Other City's Results
Jurisdictions across the country that have deployed PredPol have seen significant reductions in crime. For example, in Atlanta, GA, aggregate crime decreased by eight percent and nine percent in the two areas that first deployed PredPol in July 2013. In all four zones where the technology was not deployed, crime rates increased from one to eight percent. It was also determined that crime was 24 times more likely to occur inside an active PredPol prediction box than elsewhere in the environment over the initial deployment period, leading Atlanta to deploy citywide in November 2013.

As of March 1, 2014, the Richmond, California, Police Department saw a 21% drop in violent crime, a 28% decrease in property crime, a 50% drop in residential burglaries, and a 34% decrease in vehicle theft as compared to the same period last year.

Force Multiplier
There is nothing routine about patrolling with predictive policing. We believe that having access to this kind of analysis will serve as a force multiplier by helping to direct HPD officers where they are most likely to be needed next.