PIMS - BC Data Colloquium: Ryan Prox

The Vancouver Police Department is known worldwide for its pioneering work in the field of intelligence-led policing, and is the first police service in Canada to deploy a machine-learning predictive system directly on police mobile computers. With a focus on high-end analytics, combined with competitive technology, the Department has achieved stunning results in reducing crime rates. In the first quarter of 2016, the City of Vancouver encountered the highest number of residential burglaries ever recorded. Following the implementation of predictive technology and the consequent deployment of resources based on the forecasting, this trend reversed and the following quarter saw the lowest residential burglaries ever on record. What set this deployment apart from previous attempts at predictive policing is the way the technology was used. Innovative policing practices, combined with an advanced evaluative methodology that leveraged ArcGIS Desktop model testing and Geocortex technology, validated the system and outcomes under real-world conditions where the most successful strategies were advanced, and less successful approaches were abandoned. This learning process resulted in a highly effective crime prevention strategy that reversed a record increase in burglaries for the City.