Australian businesses failing to explore bigger data; Big data used to be purely about volume, but it is now more around the complexity of the data, which brings in notions of variety, speed and accuracy

Australian businesses failing to explore bigger data

May 12, 2014 – 10:58AM

Cynthia Karena

Externally sourced data, such as traffic flow, should be used to help businesses, Dr Hodgkinson said. Photo: Joe Armao

Despite some big data vendors claiming Australia is leading the big data analytics trend, businesses here are well behind their US counterparts.

That is the view of Ovum research director Steve Hodgkinson who addressed a Big Data conference at CeBIT last week.

Australian organisations are still only analysing their own data and are missing the business opportunities that large volumes of external data can bring to their organisation, he told IT Pro.

“CIOs [chief information officers] have neglected the external information part of their role. But more and more decisions will need to be made from analysing outside data.”

A logisitics company should, for example, analyse data from traffic flows to route transportation vehicles more efficiently, Dr Hodgkinson said. Data from smarter buildings with sensors installed throughout could reduce maintenance costs and increase safety by providing patterns of heat build-up.

Such external sources are yet to be tapped in Australia.

Furthermore, the increasing number of devices connected to the internet in the ‘Internet of Things’ will increase the amount, variety and complexity of data available to be analysed.

CeBIT’s Big Data conference explored how organisations could draw insights from all this data to make better business decisions.

Chairing the panel, Anthony Coops, business intelligence and analytics partner at KPMG said variety and accuracy of big data can’t be ignored.

Mr Coops said a car manufacturer analysed its voluminous internal data to find out what was important to customers, but when cross-checked with customer feedback found that only one of four characteristics actually mattered, with two extra characteristics identified and validated.

But getting big data projects off the ground was challenging for organisations. “By its very nature, big data analytics focuses on uncovering patterns, trends and predictions as yet unknown.”

He said this brings a degree of uncertainty that unsettles executives who want more tangible outcomes before investing in big data.

Tips

Mr Coops suggested businesses should start slow, and first identify problems that need to be solved, and if they can be solved by big data. “Then start with a pilot (or) proof of concept that is relatively easy to execute and will deliver results to help address one or more of those problems.”

Many businesses find big data overwhelming, Dr Hodgkinson said. “It requires a back-to-basics of information. How do you store it? How do you access it? How do you mine it? It’s about having internal skill development to understand how to process and analyse (increasing) external data.”

Trends

Big data used to be purely about volume, but it is now more around the complexity of the data, which brings in notions of variety, speed and accuracy as well, Mr Coops said.

Dr Maria Milosavljevic, CIO of the Australian Crime Commission, presented a case study on harnessing unstructured data at the conference, said mining comments for sentiment was a continuing challenge.

’Wonderfully horrid’ has both a positive and negative word. Is this comment positive or negative? Which one do I go with?”  A human could work this out, but the software has difficulty,” she told IT Pro.

In the future Dr Milosavljevic expects to see specialist data mining built around terms specific to each industry.

“It’s not possible to create a one-size-fits-all that is useful for all businesses.”