MAD HATTER'S
WHISPER

CodeQuantum '25 at UTSA had an Alice in Wonderland theme. Every character talks in a wildly different way, so we asked: could a model tell who's speaking?

We scraped the dialogue with BeautifulSoup, tested SVM, Naive Bayes, and Logistic Regression in scikit-learn behind a Flask API, and I built the React frontend. SVM won at 51% accuracy, and so did we.