JP Morgan data scientist on mining and machine learning

Asset management arm looks to trawl internal data for investment edge

Ravit Mandell

The use of new data might be all the rage in investing, but JP Morgan Asset Management is not solely relying on radical offerings such as satellite pictures of shopping mall carparks for a steer on economic prospects. The asset manager is also trawling through old data research from its real estate funds to get a handle on retail activity. It’s just one example of ways that JPMAM – like its peers – is looking to squeeze everything it can from what it already knows.

The firm’s three flagship real estate funds – which buy properties rather than portfolios – manage roughly $50 billion in assets globally. Over the years, the funds have naturally amassed research on the US real estate market that could now bear fruit for the rest of the business, according to Ravit Mandell, chief data scientist for JPMAM’s intelligent digital solutions division.

“Whenever [the funds] buy a mall, we collect data on the per-square-foot purchase cost, the foot traffic, when was the last time they changed the air conditioning or painted the parking lot, so we have a lot of internal data in the US,” she says.

As well as providing this data to its own investment team to steer the analysts on the real estate sector and the health of the broader US economy, the firm may give the analysis to its institutional clients to inform their broader real estate allocations.

Using data mining techniques, the firm aims to find previously unexploited uses for data held within the group.

The challenge is knowing where to look. The asset manager initially held alignment sessions to determine what intellectual property exists in the firm. “We have hired data strategists and engineers to find data internally. Someone, somewhere always has it,” Mandell says.

Mandell joined JPMAM last year from JP Morgan’s corporate and investment bank to lead the asset manager’s efforts to deliver big data and deep learning capabilities across asset and wealth management. At JP Morgan, she was head of the quantitative market making and swap trading desk where she led macro market strategy, the strategic investments portfolio, big data analytics, and market structure initiatives.

The hunt for internal data is part of a broader effort at JP Morgan Asset Management to build a data science culture at the firm, where fundamental analysts and traders incorporate machine learning and non-traditional forms of data in their daily decision making.

When it comes to data mining, the question is how to make all of that data useful
Ravit Mandell

The firm’s data efforts are housed in its so-called “data lab”, currently staffed by 21 data scientists and engineers – most of whom have been hired in the past three months. Mandell expects the headcount to double over the next 18 months.

JP Morgan is not the only firm to grasp the value of data from its banking operations. Goldman Sachs hired Matthew Rothman, the former head of quantitative equity research at Credit Suisse, earlier this year to lead a team looking to sell some of the bank’s proprietary data to clients.

Members of the data lab already meet biweekly with external data vendors. But the group’s sights are also set on applying machine learning analysis to proprietary information as well as internal data from its investment bank parent.

On the investment bank side, JPMAM is looking into applying sentiment analysis to over 100,000 research reports on companies written by JP Morgan bank analysts. The buy-side analysts are required to write reports on any company they meet, which Mandell says could provide proprietary insight for the asset manager’s investment process.

“The investment bank has been around for 150 years, and has collected so much data. As a bank-owned asset manager, we have a lot of internal data. But when it comes to data mining, the question is how to make all of that data useful,” she adds.

What’s in a word

To date, the firm has been using so-called “bag of words” language processing to analyse text in earnings call transcripts, primarily to predict revisions in earnings estimates for its emerging markets and Asia-Pacific equities, behavioural finance equities and beta strategies teams.

This method assigns sentiment scores to individual words but due to its “binary” nature, cannot handle some complexities of human language. For instance, it fails to judge the neutral sentiment of a phrase like “it’s not the worst”, Mandell says.

In the coming weeks, the asset manager is rolling out a new natural language processing model based on recurrent neural networks used by Apple’s Siri and Google. The technology is capable of memorising inputs and making predictions, as opposed to “feedforward” neural networks which are primarily used for pattern recognition. The new model will be capable of understanding the sentiment of phrases in analyst reports and other documents, in the context of the fund manager’s business.

In order to train the neural network, the firm has built JP Morgan-specific word dictionaries, which are fed into the language processing models. For example, the new models will have the ability to read a term like “higher deficits”, and determine a negative sentiment, whereas previous models would incorrectly deduce the term to be both good (higher) and bad (deficits).

Now we have the compute power to figure things out that we couldn’t two years ago
Ravit Mandell

Mandell says the new model will allow the firm to extract information from research notes, broker reports and filings to help anticipate company earnings revisions; gain insight from corporate filings (eg, 10-Ks, 10-Qs, prospectuses, etc) to enhance its research efforts; and use temporal clustering and sentiment analysis on news to better understand what’s being discussed and where consensus is forming.

For five years, JP Morgan operated a centralised data science unit that oversaw data architecture and strategy for the entire group in one place. However, the group is now diversifying to integrate data science capabilities into the four different parts of the business: central investment banking, corporate client banking, asset and wealth management, and technology.

“The math behind it has existed for many years but now we have the compute power to figure things out that we couldn’t two years ago,” Mandell says.

Mandell and her team intend to use newly developed data tools to complement the work of portfolio managers, not to replace them, mirroring efforts at other asset managers, notably Schroders.

“We are not necessarily building models to do things by themselves, but rather to augment how our investors look at things, and give them tools they may not have had prior in their day to day,” she says.

The data science team is now focused on building easy-to-roll-out applications such as clustering analysis, advanced querying of large databases like the US Securities and Exchange Commission’s website, and natural language processing tools.

Editing by Alex Krohn

Only users who have a paid subscription or are part of a corporate subscription are able to print or copy content.

To access these options, along with all other subscription benefits, please contact info@risk.net or view our subscription options here: http://subscriptions.risk.net/subscribe

You are currently unable to copy this content. Please contact info@risk.net to find out more.

You need to sign in to use this feature. If you don’t have a Risk.net account, please register for a trial.

Sign in
You are currently on corporate access.

To use this feature you will need an individual account. If you have one already please sign in.

Sign in.

Alternatively you can request an individual account here