Song Genre Classifier

This project turned out to be a fantastic lesson in the difficulty of machine learning. The goal was to classify songs into 10 genres, but the final model's accuracy topped out around 65%—far from reliable. It frequently confused similar genres like rock and metal. The low performance wasn't a failure, but a discovery: classifying something as subjective as music requires more than standard features. This project taught me more from its shortcomings than any success could have, highlighting the critical importance of feature engineering and the need for deep learning on complex, nuanced data.