First entry of the Road to Grandmaster series. You can check the acompanying blog post here. In this one we dealt with song lyrics using the 5 millions song lyrics dataset, which contains data scraped from Genius. The main objective was to use the infomation from the lyrics alone to find a way to cluster song by genre (like "pop", "rock", "rap", etc.).
In this repository we analyze only a fraction of the dataset containing 500000 songs. To download the data you can run the script data/get_data.py. It assumes you have your Kaggle API token (kaggle.json file) in the folder ~/.kaggle.
