Australian Federal Authorities Are Using Big Data To Predict Crime

Federal police look over a protester at a rally / Getty (File)

The Australian Crime Commission is now able to predict crime trends across the nation after spending $14.5 million over the last four years developing big data systems.

The ACC’s Project Fusion marks a shift into data-driven proactive law enforcement and is now at a point where the commission can start looking for trends and patterns across the nation, including scanning data for signals of emerging criminal threats.

Speaking about the possibilities big data is opening up in crime prevention at the CeBIT tech conference in Sydney this month, ACC chief information officer Maria Milosavljevic said the commission was using aspects of data mining to trawl through data sets looking for patterns and potentially predicting emerging crime issues and trends across the country.

“Being able to deal with big data is making Australia safer,” she said.

It’s not quite like the world envisioned in Minority Report – the Steven Spielberg movie where authorities can predict when individuals are going to commit crimes and arrest them beforehand – but it should identify trends that can help law enforcement agencies figure where best to direct their attention.

“It’s less about the Minority Report kind of thing which I think is too futuristic for us right now and much more about a threat that is increasing, and predicting that it is going to continue to increase based on what we’ve seen in the past,” she said.

“A lot of law enforcement work is very reactive and it’s really only where something happens in the community that we go ‘OK now we have to do something about that’ because the threat has risen to a level that’s unacceptable.

“We want to be able to be much more proactive and that is why we’re doing that.”

“One of the things we’re using text analytics for is to try to identify across the nation who could be working better together. Identify where there are different parts of the nation working on the same problems and putting them in touch with each other so reducing duplication and connecting the dots,” she said.

Milosavljevic said the agency’s crime prediction models work by integrating huge amounts of data and scanning the information for patterns, automating some of the tasks an analyst would traditionally do.

“We are using predictive analytics,” she said, adding “because our remit is serious and organised crime it’s really about looking at the threats around those elements and identifying where the threats are increasing and where we need to actually act”.

“Reducing the threat of crime in this country is really important,” she said.

“We live in an algorithmic age, we live in an age where we have access to a lot of information and we’ve moved to a world where strategy and vision setting can be adjusted on the basis of what we can see in information.”

Milosavljevic said the agency’s biggest risk is having information but not knowing what it has.

“Information is our greatest asset and our future will be determined on the basis of how well we exploit information,” she said.

(Photo by Paul Miller/Getty Images)

“If something happens and we in fact had the information available but weren’t in fact able to connect the dots then that is not good and your safety would be at risk.”

Predicting crime trends is based on identifying the likelihood of risk, Milosavljevic said.

“To be able to identify likelihood it’s about: Is there something in the data that tells me this is happening? Is there something that might be suggesting that it is? And that’s both in the structured and the unstructured data,” she said, referring to images and sounds which can be more effectively scanned automatically by machines.

“Some of the predictive stuff is about well we’ve seen this type of pattern before, it’s about the same kind of pattern that we’re seeing now but we’re really just starting that work now.”

However the commission’s remit is strategic which Milosavljevic said means it is “not necessarily predicting types of crime” but rather determining if a particular type of crime is increasing or decreasing. The analytics tools can identify relationships between entities and plot information on maps.

“It’s not necessarily about a specific crime or criminal, or a particular event taking place but its the trend is changing and that’s what I mean by strategic,” she said.

“For us what’s changing is it’s allowing us to respond more quickly, identify risks more accurately and provide information to our partners in a more timely way.”

Having the ability to process and understand a large variety of structured data including spreadsheets and some text based information is one thing but Milosavljevic said the game changer will be automating the processing and interpretation of unstructured data which includes images and sounds.

“One aspect that’s particularly difficult is the variety of data we have to deal with.”

“There are some tools that allow you to do some things but it’s limited.”