Kris Andreychuk and Stephane Contre are so in love with their jobs that you can easily imagine them poking each other occasionally, just to be sure that they’re actually getting paid to do what they do.
Recently, it took them just 40 working days to revolutionize crime data analysis to the point where they can help communities predict where crimes will be concentrated…and why!
Kris is a social worker, and supervisor of the Neighbourhood Empowerment Team that includes social workers, police officers and youth workers. Stephane is an ex-Ottawa cop who is a senior information architect on the Open City Team.
The innovative duo’s pilot ‘contextual analysis of crime’ project involved advanced algorithmic analysis of 233 kinds of data.
“Traditional crime stats look backward – what happened, where, etc. They help police deploy resources, but they don’t say WHY a crime happened, and they certainly don’t predict where future crimes might happen so you can try to prevent them beforehand,” says Stephane.
“That was our goal.”
They saw great promise in the City’s broad offering of geo-based open data, as well as other public information like census data. They believed that in all that data they could find patterns of factors pointing to either a high or low likelihood of crime in specific areas.
The data they analyzed included negatives like litter, abandoned stolen autos, liquor stores, graffiti, public drinking, bylaw infractions and pothole reports. Positive data sets included front yards in bloom, libraries, street lighting intensity, green spaces and playgrounds.
They divided Edmonton into more than 11,000 grid squares, then ran open data sets and police crime stats through a computer algorithm that looked for patterns.
The computer produced 92 ‘significant rules’ impacting crime either positively or negatively. These rules described combinations of factors found in high, moderate or low crime areas.
One such rule says crime is a higher possibility wherever:
- many stolen vehicles are abandoned
- a high level of noise complaints are received
- there’s a higher concentration of youth service organizations
- there’s a low concentration of picnic sites
Stephane and Kris can now choose a specific geographic area and generate a list of the factors tending to increase or decrease crime within that area.
Kris says the next step is community engagement.
“Data’s the beginning. Now we need to meet those who are impacted by, or can impact these rule sets, like representatives of community leagues, social workers, social agencies and business owners.
“The goal is an integrated community approach. Now we’re able to discuss many more factors – both positive and negative – that affect crime levels.
“We’ll have more flexibility to change the context within which crime does or does not occur,” says Kris.
The crime project is the first of many ways that advanced data analytics will help City employees do their work.
“Imagine, for example, applying the process to neighbourhood design. We could design them to minimize crime, but also, with further analysis, we could find factors that encourage higher participation in community leagues, or greater levels of pride in being a member of that community,” says Stephane.
Social worker Alec Stratford of the 118 Ave Neighbourhood Empowerment Team carried out the initial literature review which served as a basis for this project.