Financial Tech 2019-2020

The team is training time-series models to correlate stockand financial statement data. The end goal is to build analgorithm that will be able to predict a company’s futurerevenues, expenses, and stock price given historical data. Anaccurate algorithm would be invaluable not only forinvestors, but also for the financial planning and analysisteams, who need to accurately predict future financials forcompanies with tight margins, such as large manufacturingcompanies, to ensure a profit will be made. In particular, theproject is focused on S&P 500 manufacturing companiesbecause their historical balance sheets, income statements,cash-flow statements, and stock price are readily availableand generally clean. Before feeding data from the “FinancialModeling Prep” API into machine learning algorithms,statistical methods are used to aggregate and simplify it.Several models including linear and ARIMA have beentrained and tested on the data and their predictions havebeen compared.