This is the year when capital markets firms will have to get to grips with mountains of data; if they don’t get it under control, they will have one Big Data mess.
And Big Data can only grow. Complex sets of data arising from requirements for reporting transactions, communications surveillance, risk management (BCBS239) and swaps data repository reporting will add to the fire hose of market and trade data coming from
social media, emails, instant messages, news headlines, internal video and audio data.
This immense volume of data creates a bureaucratic nightmare for capital markets firms, few of which are prepared to handle the basic regulatory requirements, never mind take advantage of “golden nuggets” that could be useful for market monitoring and transparency.
Therefore the ability to analyse and act upon the data before it becomes stale and loses value will be a key differentiator for capital markets firms going forward.
The ability to monitor and make sense of fast Big Data hinges on connecting to multiple, disparate, live data sources. Some of these sources may be internal to your firm; some may live in the cloud or stream from sensors. You must be able to ingest and digest
all of these data streams, and able to handle the peaks and troughs of the velocity while detecting actionable patterns.
Actionable patterns can help capital markets firms detect fraud before it happens, or sniff out non-compliant trades or behavior. This becomes critical as regulators crack down, so much so that a compliance officer could go to jail for not stopping an illegal
act—it could be a rogue trader or a wild algorithm—executed by one of his
Communications surveillance is one area that poses many problems, with the need to record and analyse data from different channels such as voice, email and instant messaging. The need to retain these records for years requires a vast amount of data to be
collected. And analysis and surveillance on this data is complicated as you must look for patterns across all of the different sources, combining and correlating communications with trading activity.
In the case of fraud, it becomes more predictable when compliance officers have enough transparency to see the trail leading to the possible crime. Complex event processing has typically looked at patterns over short-ish time windows, but next generation
features and integration with large in-memory data caching technologies enable correlation and pattern detection across much longer time-windows. This opens up the technology to a wider range of problems such as money laundering, where the process of cleaning
dirty money can take time as transactions occur across multiple assets and channels to hide the trail of money and it source.
In pre-trade risk, cases such as wild- algorithms or fat finger errors can be spotted using monitoring technology which will ring alarm bells and stop the program before the market is impacted. Firms are increasingly looking at how streaming analytics can
be combined with predictive analytics and in-memory caching technologies to increase the intelligence of systems whilst retaining low latency.
Combining such technologies can help firms react to extreme market conditions and minimise losses, such as when the Swiss National Bank unpegged the Swiss franc from the euro. Not only could firms not react quickly to this stress in the market, but once
the dust settled many firms could not report on their positions or losses due to this event for several days.
A key aspect of BCBS 239 is that that risk systems should provide accurate and timely risk data under both normal and extreme or stressful market conditions. In-memory caching technologies can be vital in providing large risk data sets for calculation extremely
quickly and can also provide a much cheaper alternative to ripping out and replacing legacy systems.
These technologies will also be increasingly used on the trading floor, for example aggregating data from a myriad of databases and systems to provide real-time insight and streaming analytics to aid traders decision-making or to provide custom pricing to
clients based on historical patterns.
Rather than being buried in data, firms taking advantage of streaming and predictive analytics and in-memory caching technologies across all asset classes and systems will be the ones that unearth the golden nuggets inside it.