Natural language processing tested in the investment process through new partnership
J.P. Morgan successfully explored the use of Data Science to augment the way portfolio managers and research analysts digest information. In partnership with APG Asset Management, a Dutch pension manager, the joint team extracted data from a set of European Central Bank statements to help understand the implications and limitations of the technology. The findings are now being explored for use in other initiatives by both the trading and research teams
"The project showed how data science has the potential to augment the Research process," said J.P. Morgan's Helen Smith, Research lead for the project. "To show promising results in such a short timeframe is incredibly encouraging and we've already identified use cases across Research where this work can be applied."
Individuals collaborated via a single shared repository to enable strong cooperation between the teams. Throughout the work, the team explored the use of various methodologies from inside and outside of financial services, finishing the work by developing a prototype to demonstrate findings. Both sides are exploring additional uses and a provisional patent for the work has been filed.
Perhaps most important, the project validates the co-creation model, providing a foundation for future collaborative efforts. In addition to findings from a business context, the teams are using the project as a way to engage the data science community in different ways, share learnings and promote Research and Development in the Investment field more broadly.
"From a data science perspective we were very excited to collaborate with APG to showcase how new technologies can be used to solve real business questions," said Samik Chandarana, head of Analytics and Data Science for the CIB. "This project demonstrates how the practical application of data science techniques will help drive real value for the organization going forward."