Wednesday June 28th, 2016 - “The Contextual Analysis of Crime”
The Contextual Analysis of Crime, a pilot project in advanced data analytics aimed at fighting crime has produced a way not only to predict crime, but also to zero in on why it’s likely to happen.
The pilot program produced a method by which the Neighborhood Empowerment Team (consisting of a social worker, a police officer and a youth worker) can now predict where crime is likely to occur more frequently, and also know what factors will tend to cause the crime. Conversely, the project produced ways of showing where, and why, crime intensity will be low.
Kris Andreychuk, Supervisor of Community Safety, and Stephane Contre, Chief Analytics Officer with the Open City Team – involved taking basic police crime stats and analyzing them in conjunction with 233 other kinds of data.
About two-thirds of the data sets used – ‘negatives’ ranging from abandoned stolen vehicle locations, rowdy party complaints, litter, graffiti and proximity to liquor stores and ‘positives’ including front-yards-in-bloom, picnic sites and playgrounds – came from the City’s wide range of open data sets.
Crunching the data through a sophisticated computer algorithm, Stephane and Kris produced 92 ‘rules’, sets of multiple factors which, taken together, increased or decreased the likelihood of property or violent crime occurring within each of more than 11,000 grid squares within the city.