I recently spoke with fellow MIT Sloan entrepreneur Zeid Barakat, co-founder of Flyberry Capital, for their take on the potential of the ‘big data’ in finance. Flyberry was recently named one of two finalists in the Lion's Path Battle-Fin Tournament.
TC: Zeid, there’s a lot of talk about big data. Help us understand what it is and what the implications are?
ZB: Big data has become a pretty popular term, but its definition changes depending on who you ask. The idea is that there is an unprecedented amount of information being generated, either within companies or in the public domain. It’s been estimated that the volume of business data worldwide, across companies doubles every 1.2 years. Think of the records coming from over 1B people on the internet, and data coming from so many other sources; the twitter feeds, manufacturing data, weather patterns and medical scans, and on and on. How do companies with these large quantities of data, and make sense of this information?
Companies have struggled with these issues for a long time, but now thanks to an expanding number of tools, even smaller companies can come up with creative ways of dealing with this data. How do you store, search, or even visualize this data, and ultimately, how do you distil takeaways and map out a courses of action.
The implications are huge across multiple industries. Medical records can quickly help decode trends, crime-fighting efforts can be more targeted, manufacturers can accurately judge inventory requirements, and advertisers can target the right people.
TC: So Zeid, your company is focused on applying big data principles to finance? What does Flyberry Capital do?
ZB: Flyberry is a Cambridge startup that looks at massive varieties of data streams to find meaningful insights. Our company’s tools can be used in many ways, from helping large manufacturers manage supply shocks to identifying trading risks in the stock market. For starters we have been establishing an early-stage hedge fund, with the goal of generating returns that are uncorrelated with the stock market.
TC: Will you just be a hedge fund? How else to you plan to monetize your capabilities?
ZB: In parallel we are exploring other options for our toolkit. For example we are talking to global manufacturing firms to see if we can predict major plant failures, and are speaking to a few major investment funds to see if we can identify major event risks that might impact their portfolios. Of course, we have to be really careful about how this is done, to make sure we aren’t giving away any secret sauce that might compromise our fund returns.
TC: What does your ‘big data’ approach to investing really mean?
ZB: Well, our approach is very different from most hedge funds. While the trading strategies we use are quite common, using index and commodity futures, what is unique is the ‘big data’ approach you mentioned. We gather data from hundreds of sources, and look at things like weather patterns and twitter feeds to identify trading opportunities.
To distil this data, we rely on several MIT PhDs; experts in programming and modeling. From there, we come up with potential trade ideas, and constantly are testing and re-testing these ideas. Only the very best strategies remain, and are traded on. We’ve screened over 800 potential trade strategies and only deployed around 20 of these.
TC: Tell us a bit more about how you use unstructured data such as Twitter feeds to predict outcomes.
ZB: Good question- first we’ll process of the data stream to convert it into a dataset that we can quantify and visualize. For example, we can look for the number of free words, or combinations of words. We can also look for multiple patterns that point to the so called ‘sentiment’ of the message. From here, the analysis becomes much easier, and you can correlate those patterns to market activity.
While this sounds straightforward, we have found some data sources, such as Twitter feeds, really need other data feeds in parallel in order to be really reliable. For example, you have to manage false signals and deal with data integrity issues. Nonetheless, having such data at your fingertips can be a huge advantage.
TC: Big Data has become such an over-used buzz-word, describing advanced computation, which has been around for decades. What is the real difference here?
ZB: In a word, it’s speed. We look at it this way: in the past, traders have been mainly focused on fundamentals within a company, daily news affecting global macro economy, or on comparing companies, or currencies. With the right tools, someone can process data from the outside world much faster. Our focus isn’t on equities, it’s on events.
We’re not the only ones banking on this approach- GE recently invested $1B in a ‘big data’ research facility, and the Obama administration is allocating $200M to companies that come up with innovative techniques to manage complex datasets.
No doubt there is a lot of hype in this area, but the bottom line is that using these techniques carries a lot of promise, and we are ‘early adopters’ in finance.
TC: The hedge fund route is a This seems like a difficult time to be starting a hedge fund. With so many skeptics and angry investors out there, what makes you think you’ll succeed?
ZB: It’s definitely a tough time to be in the hedge fund industry. Overall, funds have been performing quite poorly, returning roughly 50% of the S&P, and with the news from Madoff and JP Morgan, there is a lot of distrust and skepticism out there. Obviously the increasing number of flash crashes and hedge fund explosions doesn’t help!
That said, speaking to very experienced investors gets us very encouraged that our strategy has potential. So many people are just overwhelmed trying to navigate tons of conflicting data, but an approach that uses this volatility to its advantage can hopefully offer a breath of fresh air.
Our team has also had several past successes that we can rely on; with work published on the cover of Nature as well as several other journals, and with successfully launching award winning programs on behalf of companies like Google. Our goal is to use this know-how to take major ‘shocks’ to the stock markets, and transform them into profit opportunities.
TC: Boston isn’t necessarily the first place I think of when it comes to building a fund- why settle here?
ZB: Our founding team is made up of several MIT grads, and we initially met and started collaborating here in Cambridge. While we debated heading to NYC, for us the most important asset is top talent in programming and data analysis. With great local universities and thought leaders nearby, Boston just can’t be beat. We also just love living here.
TC: How do you plan to grow from here, and are you hiring in Boston?
ZB: Our near-term focus is on building a solid track record and growing at a sustainable rate. That said, we are constantly looking for amazing people, either as programmers, modelers or quantitative analysts. Our greatest asset is, and will be our team, and we’re happy to be in a good spot to pick up some of the best and brightest!
Ted Chan is Founder and CEO of Upward Mobility and Noyo, two companies in the mobile education market. Twitter: @upwardmobility
The author is solely responsible for the content.
Meet Boston's coolest, smartest and most dynamic founders in our REEL Innovators video series!