News / Edmonton

Edmonton data shows surprising combination of factors lead to property crime

It’s like a fine cocktail.

Property crime starts with a dash of recovered stolen automobiles, a splash of noise complaints, but hold the picnic tables.

That’s the lesson the city is taking away from a new contextual analysis project, looking at all of the factors that lead to crime in Edmonton.

“There is nothing individually that is highly correlated with crime across the city. It’s the interaction of crime and environment and interaction of a multitude of different factors,” said Kris Andreychuk, who oversees the downtown neighbourhood empowerment team and helped on the analytics project.

The new project took 233 different datasets and combined all of the information to help map some of the factors that are strongly correlated with higher levels of property and personal crime.

Andreychuk said they purposely tried to cast as broad a net as possible not assuming anything was irrelevant.

“If someone thought kids eating ice-cream on a Friday was significant and they had a data set, we would include that,” he said.

Senior information architect Stephane Contre, said the key to the project was not having any biases about what might or might not be important.

“There was no subjective or bypassed selection of anything. It was the data that will tell us if it’s significant,” he said.

The datasets were brought together and then laid over a grid that divided Edmonton into 11,000 blocks each 250 square meters. When the numbers were crunched they found 92 different rules, or combinations of factors that lead to higher levels of property crime.

They included everything from the number of retirees living in the neighbourhood, to the types of business and the amount of fences in the area.

Andreychuk said it allowed them to put any assumptions aside and see what really plays into crime.

“I would never have thought that picnic sites have skin in the game as far as crime is concerned,” he said.

While police are now in the early stages of using the information, Contre said they have so much information here it can be turned onto other problems.

“Now that we have aggregated all this data and put it together, there are a lot of questions you can ask,” he said.

Big Data

  • The team developed 92 rules that correlate strongly with property crime, but stressed the correlation doesn’t necessary mean causation. Youth centres were correlated with crime, but that might be because they're located in neighbourhoods that need the social supports.
  • Another set of rules were associated with crimes against people and they said in that case it’s often fewer factors.
  • Much of the data used to create the contextual analysis of crime come from the city’s open data catalogue.

More on