Opening Cross: Can Cloud Deliver Rays of Insight from Murky Data?

Last week, we found ourselves inundated with vendors wanting to share news of new developments and client wins, ahead of this week’s World Financial Information Conference in San Francisco and the FIA show in Chicago, as companies time their announcements to maximize their exposure to important client bases. The upshot of this is that readers get an in-depth story with exclusive additional information, delivered at the same time as generic press releases hit the wires and hundreds of identical headlines flood your inbox, making it hard to know which story is a duplicate and which adds real value.

This problem also applies to traders, who want to cut through the enormous volumes of news generated each day, much of which is just noise, and much of which may be irrelevant to their particular role (more on this topic next week with news about Dow Jones FX Trader). So not only are firms using basic filters to remove the content they don’t need, they’re also using text analytics to alert them to information that might be especially important to their particular function. For example, Thomson Reuters considers these capabilities so important that it has acquired the text mining and analysis capabilities of Lexalytics (formerly Infonic), which powers its News Analytics sentiment analysis tool.

Likewise, machine-readable events data provider Selerity is also seeing demand for its content, which takes important events and decisions and turns them into a numerical value that can be processed by algorithmic trading engines, and has enlisted infrastructure provider CFN Services to provide connectivity between its existing points of presence and potential clients co-located in other locations that would benefit from direct, low-latency access to its data.

One reason it’s so important to be able to sift this wealth of information effectively is to gain real value from it, and find that “needle in a haystack”—just as investors are now doing with social media to predict market trends. Recent research claims that social media is more than 85 percent successful at predicting market movements—though that’s hardly surprising, as the same people investing in securities are likely to be those interested enough to create chatter on social media sites.

Of course, to get the most value out of any new information, you need a store of historical data for comparison, as Rich Brown of Thomson Reuters says—especially if you have algorithmic traders wanting to incorporate a dataset or venue into their strategies. For example, earlier this year, Direct Edge began receiving demand for historical data on its (still-young) market from algo traders wanting to build and evolve their strategies, and enlisted cloud data specialist Xignite to build and host a storefront for its data.

But the need for new approaches to storage and delivery doesn’t just apply to news, sentiment or historical data. For example, Marshall Wace is using ITRS’ Geneos tool to capture, store and analyze historical capacity metrics to assist in future capacity planning decisions—an area where some firms are now turning to cloud computing environments to mitigate the impact of deploying high-performance applications, such as Celoxica, which is making its new Container combined data and trading appliance available as a managed service via Options IT’s hosted infrastructure, to make it easier for clients to test and begin using the product. Making it easy to test and use new tools was also cited by Selerity CEO Ryan Terpstra in his dealings with CFN, and—considering the lengthy and rigorous testing procedures performed by OneMarketData’s latest clients, Galiam Capital and Scottrade, a key area for cloud services to prove themselves will be in helping other vendors to prove themselves, by enabling ready test environments.

For tips on how exchanges can get the most out of the cloud, check out Ross Inglis’ Open Platform. Certainly you won’t need a text analysis engine to find the value in that.