Banned Books Project

MadData 2025 Submission

Chieler Li, Steve Lin, Aniruddh Mayya, Sanjay Murali

Libraries

pandas
streamlit
folium
scikit-learn

Inspiration

We were inspired by the mission of PEN America in fighting book bans in the US.

Functionality

The program takes user input for title and author and uses a model to predict the likelihood of a book being banned, based on title, description, and genre. The model is trained on data scraped from GoodReads.

Methodology

We scraped data on GoodReads about banned and non-banned books, and trained a sci-kit Learn XGBoost classifier to classify book as likely to be banned based on the aforementioned features. The user inputs a title and author and we scrape GoodReads to get it's information, then use our model to classify it.

Challenges

We found that scraping took a large portion of our time, as the number of books we chose was very large. Of course, the 24 hour time constraint was also a challenge.