Preparing for a Data Science Career in Finance with a Master’s Degree

Sponsored School Search


In early 2012, hedge fund managers started to notice a curious deviation in a certain obscure derivative investment fund. The Markit CDX North America Investment Grade Series 9 10-Year Index fund – better known as CDX IG 9 – was losing value for no apparent reason.

With further investigation, it became obvious that someone, still anonymous at that point, was taking massive short positions against the index. Insiders dubbed this unknown trader “The London Whale” for the location where the trades were booked and for the sheer size of the position.

Many of them took the other side of the trade, playing their own analysis and going long. They won.

The London Whale, eventually revealed as J.P. Morgan trader Bruno Iksil, wound up costing his firm somewhere between $7 billion and $9 billion.

Quants Own The Street … and the Psychiatrist’s Couch

The most talented “quants” – trader shorthand for quantitative analysts – occupy a rarified niche in investment banks and brokerage firms thanks to their knack for working out predictive models within a bottomless sea of numbers – models with the potential to generate enormous profits.

Today’s quants see beyond the dry fodder of profit and loss and risk assessment, with Cutting-edge firms like MarketPsych now delving into what is being called “behavioral finance.” This emerging field seeks to combine cognitive psychological theory with market data to explain and forecast irrational financial decisions.

In 2011, Derwent Capital, a London-based investment management firm, launched the first-ever Twitter-based hedge fund. Their algorithm monitored and analyzed tweets in real time to predict stock market behavior, and it did so successfully: The experiment ran for just one month, and in that short 30-day timeframe the average return was 1.86 percent, with investment predictions hitting at an astonishing 88 percent success rate–substantially outperforming the market.

Derwent’s experimental algorithm was unique in that it used Twitter data exclusively. Subtler, and less highly-publicized, algorithms incorporate social media cues, interweaving them with more traditional analytics. Finding the right blend is an exercise of almost endless fascination for data science purists.

Behavioral finance is being used for good as well as profit. A 2014 article in Bloomberg News discusses ways that governments and consumer-oriented application developers are using the theory to help average consumers make better financial decisions in their daily lives.

Wall Street’s Need for Speed: The Quest to Shrink the OODA Loop

Trading stocks has always happened in more or less real time. It’s just that the definition of real time has changed, from the seconds required to exchange hand signals across a trading pit, down to the milliseconds in which automated trading algorithms handshake on a deal deep in the circuitry of networked computers far from the trading floor.

It’s an advance that has been a long time coming, and that shows no signs of fading away. In 1867, a Mr. E.A. Calahan of the American Telegraph Company invented the first stock ticker machine. Spitting out a constant stream of symbols and prices on ticker tape, it was a boon to investors of the era; anywhere in the country, they could monitor and act on the price of their investments in as little time as it took to fire a telegram back to New York.

Every advance in trading data since then has exhibited a similar quest for speed, a shortening of the OODA (Observe, Orient, Decide, Act) loop that would allow traders to outmaneuver one another and profit from the ability to act rapidly on new information.

From The Theory of Speculation to High Frequency Trading

Quantitative financial analysis isn’t particularly new. Many trace the field to a doctoral dissertation published in 1900, Louis Bachelier’s “The Theory of Speculation.” The paper discussed the application of Brownian motion – or the stochastic process of particle motion – to options pricing. It was the first application of advanced mathematics in the study of finance and the beginning of what would come to be known as quantitative financial analysis.

What is new is the speed at which quantitative analysts can make an impact when trading. Trading systems based on mathematical rules emerged as early as the 1980s, but the primitive state of computer systems at the time meant that these formulas were applied manually and still had to be closely monitored by human eyes.

By the late 1990s that all changed when the Globex electronic exchange had been instituted in Chicago. Computers, which are, if nothing else, good at math, could now be programmed with rules and taught to execute trades themselves, in milliseconds.

The groundwork for high-frequency trading had been laid.

High-frequency trading is not simply about issuing buy or sell orders quickly, although that is an important component of the strategy. Rather, it is about issuing many such orders sequentially—buying and selling in fractions of a second, after market movement of as little as one cent. Enough of these orders issued quickly enough can run up millions of dollars in profits – in minutes.

No human is necessary in this process… except for the data scientist who devises and deploys the algorithms the system will execute.

With Great Speed Comes Great Risk: Data Science in Financial Risk Assessment

As any trader will tell you, managing risk is the real work of financial institutions. Data scientists in finance analyze risk from all angles… everything from political unrest to environmental concerns to supplier instability.

This involves crunching wide swaths of data from diverse sources. Often, this overlaps with other sectors that have come to rely on talented data scientists: environmental risks are understood and informed by incorporating data generated by environmental monitoring; health concerns are assessed with input from health sciences databases.

You Gotta Know When to Hold ‘Em

When you are going to be launching 10,000 orders per second, you need to know that you have a statistical advantage in doing so. Such trades play out too fast for human intervention, and sometimes rely on sequences that must be executed in their entirety – or not at all. The restraint of the gambler takes on superhuman dimensions when combined with the cold analysis of a trading algorithm.

Even so – in a market with other high-frequency traders using comparable technology, quants know to expect the unexpected and never stay off their toes for long.

On May 6, 2010, in the span of the half-hour between 2:30 p.m. and 3 p.m. the Dow Jones index both lost, and regained, nearly $1 trillion. The sudden plunge was almost entirely the work of interacting high-frequency trading algorithms, responding to one another in ways their creators had never imagined.

An Industry Run On Numbers: Data Science Makes its Home on the Street

As an industry whose chief concern is numbers, it’s no surprise that the financial services sector is a natural fit for data science, providing no shortage of opportunities for master’s-educated data scientists in a variety of roles to get in on the action:

  • Front office quantitative analysts working to determine fair stock prices, reduce trading risk, and identify profitable trades
  • Asset management analysts using data analysis to develop investment strategies for asset growth and protection
  • Algorithmic trading analysts developing trading algorithms to beat average market rates of return based on improved data or improved interpretation
  • Corporate finance experts analyzing risk and prospects for business investments by banks

In each case, only the best and the brightest need apply, creating a golden niche for those with master’s degrees in data science. Enormous sums are involved and the stability of corporations or even nations can be affected in the course of trading at the levels that have become common in modern finance… just ask the London Whale.

Back to Top