Analytics helps Edmonton police predict crime trends

The Edmonton Police Service (EPS) hopes an analytics tool from IBM will help the agency to not just report on past crime trends but also forecast and predict them.

The police force’s phased business intelligence (BI) program began in 2005 and experienced some early gains.

Today officials and front line personnel can extract near real-time data from the system – and use that to produce trend reports.

But with the new forecasting capabilities – that kicks in sometime next year – the police service expects to get even more benefits from the system.

“Right now we can see where and when crime is happening and that helps us plan for better force deployment,” said John Warden, BI project lead for the EPS. “But the ultimate goal is enhanced forecasting and prediction.”

Forecasting and prediction, however, do not involve some science fiction movie-like capability to determine that a particular crime will occur at a given time and place, Warden noted.

Rather, the police service expects that by next year its business intelligence system would offer insights into such questions as:

  • Which kinds of people and groups are committing crimes
  • How are crimes being committed?
  • What are the motives?

By inputting answers to these types of questions, the analytics tools would help the Edmonton police come up with trending models.

These would shed light on: times of the year certain crimes are more likely to happen; or conditions conducive to those crimes.

Analytics is already widely used in the private sector to help companies determine business trends.

It helps firms develop business strategies and campaigns, said Warden. “These tools can also assist law enforcers in crafting forward-looking rather than reactionary strategies.”

For instance, with such information, the force can better determine when to beef personnel deployment to a certain location or pinpoint the best combination of policing tactics to use in a certain situation.

At its current phase, the Edmonton police’s system has reduced the force’s crime reporting and repose time.

An automated incident reporting procedure, which involves police officers entering crime scene details on to a networked computer in a squad car, is pretty normal for almost every police force in Canada.

But the Edmonton police force’s analytics system takes the process further by crunching crime statistics in the backend and using algorithms to detect patterns.

The results are crime trend reports available to top level municipal and police force officials in days rather than weeks or months.

Traditionally data would have been collected from hard copy reports and then evaluated by crime specialists and analysts and recommendations relayed to the police chief after several weeks, said Warden.

“Now we can determine a crime spike in a given area, have that report delivered online to city and police officials the moment it’s finished and the force can have a team working on it the same day.”

The whole system depends on being able to provide real-time data, says Kim Devooght is the vice-president, public sector at IBM Canada.“Relevant and timely data is fed into the system produces better and timely insight for stakeholders and planners.”

The use of analytics in government is also an emerging trend, Devooght said.

Government organizations generate huge volumes each day.

There’s a need for tools that can make sense of this information quickly and in the most meaningful way, he said.

Business analytics is being used in crime reporting and prevention in other large police forces such as the New York Police Department and the Miami-Dade Police Department, Devooght said.

Share on LinkedIn Share with Google+