Big data in the real world

Fighting fraud

While big data lends itself very well to long-term research endeavours, the near-real time analysis of large pools of information also has significant value in more short-term initiatives – particularly when it comes to spotting outliers in consumer behaviour.

This becomes especially relevant when it comes to fraud detection in the financial sector. Fraud can cost banks millions each year and often leaves victims in a very difficult position until they can get back on their feet. With the ability to trend data on transactions against where, how, and when customers usually go about their banking activity, banks can develop better internal processes to detect fraud and take proactive action before things get out of hand.

The UK Cards Association recently revealed that the UK is Europe's leading online shopping economy, with spending by British consumers rising 16% over the past year to £91 billion. The convenience of card payments represents the main driver for this growth. Unsurprisingly, the value of fraud losses on UK cards also rose 16% in the same period.

In response, banks and financial regulators have already begun to use rapid large-scale data analysis in their fraud prevention efforts. This allows them to spot suspicious spending before customers even take notice so they can proactively block peoples' cards before fraudulent activity gets out of hand.

Of course, blocking customers' accounts frequently can also begin to take its toll on banks' relationships with the people they serve. The ability to distinguish between suspicious spending and the odd splurge on a new television or holiday adds a complex dimension to fraud detection, which is why the "whole picture" approach afforded to banks by big data analysis is well-suited to this task.

The practice of applying big data to fraud detection has also gained traction in financial markets. The Financial Conduct Authority collects 13 million daily transaction reports from regulated financial institutions across the UK and Europe, and requires fast, timely access to detailed intelligence drawn from this data set. With its big data tools, the organisation can complete its analyses in a matter of seconds, which allows it to quickly spot and act on unusual trading activity.

Breaking down the winning formula

Big data analysis has even found its way into the sporting world. When it comes to professional football, coaches are always looking for new ways to help their teams gain a competitive edge. Even the most experienced leader struggles to keep an eye on all of his players in the heat of a match; much less take stock of everyone's positioning and passing accuracy for a full 90 minutes.

Applying analytics to data captured using motion capture technology, for example, coaches can visualise and evaluate players' performance over the course of a game (or season) and pinpoint the skills each player needs to build on to help the team improve. Match data can also serve as a useful tool in recruiting new talent, offering coaches and managers an in-depth view of a prospect's playing style and performance over extended periods of time. This makes it easier for them to identify and attract players that will best complement their team.

When it comes to team sports such as football, the value of data analysis chiefly comes into play off the field. In the case of professional sailing, however, this technology has become an integral part of the action as it unfolds.

At the America's Cup last autumn, Oracle Team USA completed what was arguably the most thrilling comeback in the history of the event. Trailing New Zealand eight to one, the Americans bounced back to win nine races in a row and secure the title.

Housed behind the scenes in the hull of the winning team's 72-foot catamaran was a server collecting live data from the 300 sensors installed all over the boat. The sensors were programmed to record 3000 variables at a rate of 10 times per second, which were in turn analysed by a four-man crew to help Oracle Team USA make the intelligent split-second decisions that ultimately carried them to victory.

A bit of future gazing

Unsurprisingly, it does become difficult to discuss big data without imagining how it will evolve in the coming years. Across many applications, some obvious and some less so, the leaps in speed and scale that big data analysis brings to the table will encourage people to re-evaluate how they approach the challenges they face. In the fields described above – medical science, research, finance, and sport – the opportunities afforded by large-scale data analysis have already begun to have a significant impact.

In truth, big data analysis is still in its early stages. People have started to link information from a large number of sources in ways that have never before been explored, but there is a lot more insight to be uncovered in the mass of unstructured data being collected today. These "known unknowns" – potential revelations hidden in the uncharted territory between structured data points – will breed an increasingly more judicious decision-making culture as they come to light.

The fact that we don't know exactly what these discoveries are going to look like may upset the big data sceptics, but in reality this level of uncertainty and the potentially unlimited range of applications for big data are perhaps the most exciting things about the technology. If the progress we've made so far is any indication of things to come, the future looks bright indeed.

  • Dermot O'Kelly is Senior Vice President at Oracle, responsible for driving Oracle's operations and growth across the UK, Ireland and Israel. As such he has a unique view of how new technologies are changing the way people work in these regions.