The problem Profit Prophet solves

Many young people find finance daunting and are hesitant to start investing. Profit Prophet is a data-driven stock prediction game, designed to help build interest in the financial markets.

With Profit Prophet, users are tasked with predicting the closing price of a stock relative to its open price given a same-day news headline. Once the user submits their guess, the stock chart reveals itself and the user receives corresponding feedback. The user then has the option to seek AI analysis via GPT-3.5 to help interpret the stock's movement given the headline. This feedback allows users to gradually understand the dynamic between major events and the stock market. The user can complete further rounds in rapid succession, each time earning points relative to a virtual opponent.

Ultimately, we hope Profit Prophet can help engage all individuals to learn about the stock market. Investing is of central importance to individuals of all ages, and we hope our game can be an entertaining spark toward greater financial literacy.

Challenges we ran into

To encourage the user to keep playing, we wanted the user to compete with a bot. We considered two classification models to predict price movement, an LSTM and BERT. However, we encountered difficulties in using BERT as Hugging Face was in maintenance on Saturday. Nonetheless, we implement an LSTM to predict the percent changes in the stock given the headline.