After running a series of consultation workshops with Australia’s defence and law enforcement agencies, the $80 million CRC has drawn up a five-year research roadmap for its data analytics projects.
These include using data streams to build a Wikipedia-style briefing resource on criminal activity, data privacy protection policies, and integrating different datasets across national and federal law agencies.
The CRC’s chief technical officer Dr Brenton Cooper said building machine learning, or “machine enablement”, is a critical component of data analytics. Sophisticated machines can collate and scan a vast volume of material, and are programmed to pick out key phrases, figures and spikes in social media activity that could be relevant to counterterrorism operations. The information will be used to build digital technology tools for defence.
“We’re building an app called Beat the News,” he said. “The idea is to develop a warning system based on data from a wide range of freely available sources that can map social responses to things such as food prices, cost-of-living pressures, crime rates and local news events.”
The app is being designed as a data analytics tool for defence strategists, and Cooper explained that the system is focused on mapping “population-level events” as reflected by social media patterns, rather than individual use. “We’re not going to be interested in what Joe Bloggs is doing,” Cooper said.
The CRC is also working on a project to build a rapid-response briefing tool that will collate data and present a Wikipedia-style page of information on an emerging threat. Cooper uses the example of a ship that might be suspected of smuggling drugs into Australian waters.
“We’re working on a system that could rapidly pull together all the information that’s needed on that particular ship – where its last port was, where it went on its most recent voyages, and whether any of those ports are implicated in global drug smuggling operations,” he explained.
“Instead of being swamped with information options – which is what happens when you use Google to find something – we’re building a tool that will provide analysts with the information they need, quickly and efficiently.”
Another data issue facing Australia’s police forces, and other law enforcement agencies such as customs, is the lack of a central data repository. Can state and federal data sets be combined? It’s not as simple as it sounds.
“It’s a complex and sensitive area of data management,” Cooper said. “There are questions to be resolved around data ownership, access and responsibility for maintaining a centralised data repository.”
Privacy is also a key research area, as is public education about how data analytics can be used to benefit society. The Cronulla race riots that occurred in Sydney in 2005 predate Twitter by just a year, and Cooper said it’s possible that data analytics of a spike in Twitter activity (had the mini-blog site been around) would have predicted that tensions were likely to erupt.
“People might be uneasy about data analysis of social media activity, but we’re looking at patterns not individuals. It’s a bigger social picture.”
– Rosslyn Beeby