United StatesPublic sector2011–2020Los Angeles Police Department
PredPol geographic crime prediction
LAPD ended PredPol after an audit could not establish its crime-reduction effect
Problem
LAPD wanted to direct limited patrol resources to places where vehicle-related crimes were expected, reducing crime and victimization.
Approach
PredPol began in the Foothill Area in 2011 and used historical crime types, locations and times to generate patrol hotspots. A 2019 inspector-general review found inconsistent patrol-duration records that prevented a clear assessment of effectiveness. In April 2020, Chief Moore cited pandemic budget pressure when announcing cancellation. The National Institute of Justice subsequently confirmed that LAPD no longer had a PredPol contract.
Results
No published figures
Takeaway
The department could generate hotspot maps but could not reliably reconcile records of the patrol activity intended to make those predictions useful. That left its operational benefit difficult to establish.
Categories
- Size
- Large enterprise
- Industry
- Public sector
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Partner co-development
- Evidence
- Regulator, audit or court
Sources
Source summaries
The review documented PredPol's 2011 introduction, intended use and inconsistent measures of time spent in predicted hotspots. Those discrepancies limited conclusions about crime reduction.
Moore announced cancellation. He defended PredPol.
The profile described the predictive-policing model and reviewed experimental evidence about it. Its implementation section confirmed that LAPD no longer held a PredPol contract.
Full details are in the original sources.