Jupyter Notebook based interface to download and visualise music artists discography data, statistics and lyrics sentiment

This project explores methods to classify song lyrics by genre across two notebooks. In Notebook 1, various models (Transformer, LSTM, Random Forest) were tested, achieving accuracies up to 93.5%. Notebook 2 investigated Graph Neural Networks (GNNs) with TF-IDF and BERT embeddings, with the highest accuracy of 85.55% using GCN and BERT embeddings.