PapaQuant

PapaQuant

Inspiration

We see millionaires be dumb and say things that ruin their reputation in the national media all the time. We saw this as an opportunity to profit off of their mistakes. The primary inspiration for this project, and its namesake, is Papa John, who ruined his reputation and the reputation of the popular pizza restaurant in one afternoon.

What it does

Our project scrapes popular business sites, like Forbes, and searches for potential ceo controversies. It then tries to classify them as potential long term growth or regression. It will explain whether the best idea is to either short or buy the dip after major controversies.

How we built it

We discussed a pipeline of first classifying major controversies, then it summarizes the articles using ChatGPT, then it does a sentiment analysis on the summary to decide if it is sarcasm or a joke, and if it identifies as a real controversy, then it will go through our XGBoost model to decide if this is a long term or short term failure.

Challenges we ran into

Scraping was very difficult as we kept getting blocked from major sites. We got around this by scraping wayback machine instead of the sites themselves. We also struggled finding accurate, updated datasets to train our model on.

Accomplishments that we're proud of

We all worked very well as a team, and worked well off of each other. Each of us had our own skill sets and we were able to build quickly and separate work well. We also were proud of how well our pipeline actually worked, we were able to get it working very well. We spent a lot of time on the frontend and getting the UI in a place that we liked a lot.

What we learned

We learned that this is actually a very well researched field, making our development a lot easier having other research to pull from. We also had limited experience with sentiment analysis, so we learned a lot about how that is calculated and found out. We've built machine learning pipelines before, but this is one of the more advanced ones we've built before.

What's next for PapaQuant

We would like to make it an active trading platform. Eventually with more research and resources it could potentially have access to automatically trade with the money given, doing low amounts of money for high risk and high amounts of money for low risk.