APFIC Panelists: Big Data Presents Big Challenges in Asia

While panelists agreed that analyzing large volumes of public unstructured data─such as social media and sentiment data─as well as the ability to apply Big Data analysis to internal datasets are both valuable, they disagreed on where these techniques can be put to best use, and how they can be monetized to deliver a trading advantage.

"Over the last two decades, banks have invested a lot in structured data. And over the next few years, we'll be figuring out how we deal with unstructured data," said Shiv Kumar, global head of cash equities technology at Macquarie. "We've started flirting with... how we can look at storing this unstructured data and analyzing trading patterns. It's all absolutely invaluable. It's just that we haven't been able to structure it into a commercial proposition that we can use in our algorithmic trading engines."

Another challenge to using Big Data in this way is the challenge of accurately and consistently analyzing different languages. "We haven't moved as far as news feeds and sentiment engines," said Catherine Turley, head of analytics at CIMB. "It's not just about different languages, but different cultural interpretations in Asia... and even if you make this machine-readable, it is harder for a machine to understand these nuances." And with so many different jurisdictions across the region, there are no common standards about how information is published, she added.

Others suggested the way to see quick return on investment will be to apply Big Data analysis techniques to established datasets, rather than trying to figure out how to derive value from new data types. "If you want your algos to react to news or Twitter feeds, you may not see an immediate benefit. But if you apply that to data you already have... you may see quicker ROI," said Vinay Pampapathi, equity business manager for Asia at Daiwa Capital Markets.

Indeed, firms may have most success using these techniques for displaying large datasets to understand correlations and identify signals that might be otherwise lost in the volume of data. "We're already there for real-time visualization of Big Data. You can't rely on charts or graphs in spreadsheets: you need to be able to visualize data quickly and take action," said Scott Southward, director of products and services in Asia Pacific at Datawatch. "The only reason more people aren't using this is because they don't have the tools to hand."