This repository contains our work on Data Science over the Spotify Dataset. The idea is too predict the genre of a music and its popularity to determine the future hits. For the first part, we used GradientBoost to predict with a f1-score of almost 0.7 . For the second part, we used RandomForest. However, a feature was bad quality so we had to use method to increase the precision : KNN, Word Analysis and Clustering. With this last one, we got a R2 score of 0.56.
Made with Quentin Parent.