Song Popularity Prediction and Personalized Recommender System
This is a Machine Learning project on the Spotify Million Playlist Dataset. The purpose of this project was to solve two business problems:
- Predict song’s popularity based on song’s audio features.
- Design a personalized song recommender system based on song’s audio features.
The first part of this project involved usage of Supervised Learning methods whereas the second part was focused towards Unsupervised learning methods.
For the first part, we used methods such as Logistic Regression, Bayesian Logistic Regression, Decision Trees, Random Forest, KNN and SVM. Whereas, for the second part, we used K-means and Hierarchical clustering and Mahalanobis distance.
Please feel free to take a look at the Jupyter Notebook for more details and code!