The (accidental) art of making Big Data analytics profitable -CSIRO

CSIRO has found itself producing products a long way away from their original expected uses, and its going to make the world a better place.

CSIRO’s director of business development – information services, Geof Heydon, said that in our modern digital economy, one that is defined by large data sets, that the key differentiator is analytics.

It has seen Australia’s national science agency produce a wildly varying array of technologies designed for one purpose, being reused for others – with great success.

The problem Heydon believes with a lot of businesses trying to make sense of their datasets is the large amounts of marketing overhype in the IT sector, specifically in the areas of cybersecurity, the Internet of Things (IoT), Big Data and in communications infrastructure.

Heydon said all these areas are rapidly being commoditised, and the one thing they all share in common, and can produce the most benefit is the analytics of the data produced.

Consumers don’t really care whether their Internet comes via Vodafone or Telstra, which app does what, as long as it all works. Our communications infrastructure is already commoditised, he says, noting that each company struggles to differentiate their offering in the marketplace. We only tend to notice when a brand lets us down.

“We only tend to think about our roads when they’re congested – same with broadband,” he says.

One of the key drivers of the 21st century will be the Internet of Things (IoT), pretty soon every digital piece of technology will be online through this commoditised infrastructure.

While Heydon admits, the idea of ‘the connected fridge’ seems absurd, it is the future unforeseen uses of the technology – especially in combination that will define it. An example being the dishwasher – a combination of electricity, sewerage and water, which no one foresaw in a world where neither of these infrastructural devices was readily available to the masses – let alone in concert.

Heydon said one day you will be able to buy a ‘washing cycle’ (as a service effectively), without having to buy the whole machine – the sophistication of robotics and sensors will have reached that point.

The usage of sensors also brings privacy concerns, as does Cloud and Big Data, its storage and analysis, only when all these issues are settled can these combined tech innovations really begin. The problem is that the rate of innovation is vastly exceeding the rate of policy, that it is ‘constantly moving on muddy ground’ when it comes to laws, says Heydon.

However, Heydon said data analytics could be potentially a huge opportunity for Australia. Its an export that can transition to anywhere. However, Government policy in encouraging this in the digital age is lacking – its the “difference between an export to somewhere, versus an export to everywhere.”

“If we’re not careful, we’ll end up as a net importer of everything,” he said.

One of the hardest problems Heydon faces is attempting to get various government agencies to work together – especially since in the future they will not be so distinct.

One of CSIRO’s most high profile projects has been the Zebedee – a 3D environment scanner. It was originally used to map changes in the environment, but now it is being used by Queensland police to map crime scenes quickly.

For example, after a crash on the motorway, officers can quickly 3D scan the entire scene, gather all the information they need, and quickly reopen the motorway – instead of shutting it for hours and causing a traffic jam. These are the unforeseen consequences of technology, and data analysis. The cost benefits, in both time and money, reduce wastage.

From here, the technology has been used to 3D map the exterior and interior of the Leaning Tower of Pisa – producing the first ever accurate rendering of the building, which the Italian government can then use to produce solutions to keep it from sinking.

Airline companies have already signed up to use the product to map the interiors of newly built jumbo jets, which can then be kept on file to monitor changes to the aircraft structure over time – such as cracks or in the aftermath of an accident.

CSIRO’s agricultural sensors have also been adapted to plot the course of bushfires, ensuring prevention and timely reaction from emergency services.

It has also developed new systems to collate and measure agricultural data, using less than conventional means. In Tasmania it is studying the data from oysters to understand the effects of chemical runoff from agriculture, and studying grapes to predict botrytis – and tracing where the pesticide run off goes, and how it affects soil fertility. Both of these methods have been exported to external markets.

Another project that saw unforeseen consequences was CSIRO’s analysis of a database of two billion tweets. Asking theoretical questions such as ‘what is normal?’ and what happens to tweets that are outside the ‘normal’ parameters also leads to being able to track the source ‘tweet’. This was demonstrated when CSIRO knew about the Christchurch earthquake 20 minutes before the New Zealand media did.

Other digital tracking and data analysis systems, initially developed for sports scientists, are now used by the major TV networks to track player performance in game, and in app. Viewers can track whether players were out of position and how many steps they take in a game, a profitable new set of statistics to exploit. As an aside, there has been interest from other industries, such as mining, for emergency preventative measures.

“All of this is really just the Internet of Things going crazy with analytics on top,” Heydon said.