Victoria Police eye data analytics to reduce criminal informant risk

Applying algorithms could help force decide who to recruit as a 'human source' and when

Despite the extensive digital footprints left behind by even the most cautious criminals, sometimes social media records, IP addresses and metadata offer police scant clues about planned crimes.

“How do we stop or prevent a crime actually occurring when we don’t have digital assistance? One of the best ways we can do it is to recruit a human source,” explains Mark Rollo, Acting Superintendent at Victoria Police. “It’s about recruiting the right people to put them into positions where we can’t get technology into.”

For many years informants have been recruited within criminal groups, from motorcycle gangs to terror cells, money laundering operations and paedophile rings. Recruiting insiders has always been incredibly risky work for the officers involved.

To reduce the risk, and help the officers target just the right people at just the right time, Victoria Police are considering the application of advanced data analytics and algorithms to its multiple data sets.

“It’s about getting that information from the right people,” Rollo says.

Reducing risks

Selecting potential informants until recently had been very manual work, reliant on paper-based process and the experience and intuition of officers.

That could be about to change as part of Victoria Police’s wider deal with SAS which will result in a new software platform to allow the force’s crime analysts to query and integrate results from discrete police databases.

Victoria Police went to market in mid-2016 seeking a solution that would integrate its array of datasets and “provide enhanced capability to search, validate and analyse data” and deliver “enhanced information and predictive analytics, particularly for high-risk police activities such as family violence and counter terrorism”.

The project is being funded through the $227 million set aside for IT upgrades for the force in the latest state budget.

The SAS solution will pull in data from LEAP – Law Enforcement Assistance Program – a custom-built mainframe application for incident and offender tracking, theft information, vehicle information and crime reporting; Interpose, the force’s principle intelligence database; the iFace (facial recognition) database; MACPAC (MDN and CAD Patrol Activity Collation); the Sex Offenders Registry and National Child Offender System; the Licensing and Registration System (LARS); the Traffic Incident System (TIS) as well as from social media and other public sources.

As part of the solution, the application of algorithms may be used in the future to score potential recruits, helping pinpoint who best to approach and when, Victoria Police said.

The solution could also help police map the connections between individuals in criminal networks.

“The use of network analysis may be able to show who is connected to whom. Does that person have access to the right people, the people we’re trying to target?” Rollo explained.

Recruiting informants is a high risk activity but invaluable in thwarting crime. Data analytics could help remove some of the risk, Rollo says.

“How do I de-risk going to approach someone so they can provide information? How do I de-risk that for the community? We may be able to use analytics over the top of that data to do that even better.”