- Barclays data science research group is still fairly new - having only been established in 2018 and publishing reports consistently since 2019.
- Ryan Preclaw and Adam Kelleher, who co-lead the group, spoke to Business Insider about how they adapted their approach over recent months to stay on top of the coronavirus.
- Preclaw said Barclays' data-science research group was forced to react to data sets that deteriorated in quality or usefulness in order to stay ahead of the curve.
- In recent weeks it has published multiple reports and high-frequency indicators, which give close to real-time analysis via alternative data sets.
- Click here for more BI Prime stories.
As market volatility reaches all-time highs thanks to the spread of the novel coronavirus, banks are looking for any edge they can get when it comes to insights on how financial markets might react next.
The unprecedented situation has created an opportunity for banks' research and data-science teams to demonstrate the capabilities they've built. One example is Barclays, which is just two years into establishing its data-science research group.
Ryan Preclaw, who co-leads the group, told Business Insider the bank was able to cut down a response time that might have typically taken days or weeks to get out analysis on the impact of coronavirus.
"Because we have a lot of the tools already built now, we're able to go in and take a good idea very quickly. Look to see if it's going to work and be helpful. Once we see that it is, we can get something published out to clients in very short order," he said.
Barclays' data-science research group is made up of two teams: investment sciences and data science. Preclaw heads up the former team, which focuses on implementing data science into the investment research process. Meanwhile, Adam Kelleher leads the latter team, which targets building analysis and models on top of data sets brought into the bank.
And while the group has been in place since the beginning of 2018 and publishing reports regularly since 2019, the impact of the coronavirus on financial markets has forced an adjustment in strategy. In recent weeks it has published multiple reports and high-frequency indicators, which give close to real-time analysis via alternative data sets such as public or geolocation data.
"It's definitely the most extreme or most dramatic opportunity," Preclaw said. "Prior to this, a lot of the value was in the flexibility of what we could do. That we can take a look at a situation that people were otherwise having a hard time understanding how they we going to get the nuances and figure out what's going on. This has allowed us to demonstrate that we're able to add speed on top of that."
Barclays adapted its approach in recent months
Preclaw said interest in coronavirus first came across the group's desk a few months ago. Initial work was based on data coming out of China. The goal wasn't to make predictions, he added, but instead to give people a framework so when new information was available it could quickly be interpreted.
Eventually, however, a change in strategy was required as questions around the quality of the information were raised, Preclaw said.
"If we're really uncertain about the process by which that data is being generated, it's harder for us to feel confident about the kinds of insights that we're drawing from it," Preclaw said.
At the same time, things were accelerating in the United States. As a result, the team looked to more high-frequency data that was updated more often and would provide closer to a real-time look at things. The goal was to understand how government policies promoting social distancing would impact businesses.
Looking at open data that was publicly available, such as subway ridership or hits to the CDC website, proved fruitful for a while, Preclaw said.
However, eventually some of that information became obsolete. As Preclaw put it, you didn't need a data feed to understand no one was riding the subway anymore.
Once again changes were required, with this time more focus on geolocation data, which provides insight into the location of electronic devices, such as cellphones, to understand how effective policies are at keeping people at home.
"We can do it with data that is a little bit more flexible and gives us a little bit more of a breadth of different things that we can look at to understand how it will change it," Preclaw said of using geolocation data.
Preclaw declined to get into specifics around types of data Barclays would be looking at going forward. However, he did say that clients are still only interested in big-picture, macro trends for the time being. The general sentiment: How much are economies slowly down, and at what pace?
Eventually, focus will shift to understanding when things have bottomed out. That, in turn, will lead to people asking about specific segments, Preclaw added.
"We needed to do some fast stuff to let people have a view into the world when it was unfolding very quickly," Preclaw said. "I think now the direction is to take a little bit of a step back, try to find the deeper insights that are going to take a little bit more work."