Team members: Jade (Yiyang) Cao, Betty (Yi) Chen, Cindy (Yuexin) Chen, and Michelle (Ruoxuan) Liu
Spotify’s track dataset on Kaggle contains diverse information about songs such as their artists, popularity, and genres. Significant business values can be generated if the dataset is used to encourage users to explore new songs similar to their current interest. To achieve this, we first performed Principle Component Analysis (PCA) and Clustering Analysis to identify groups of similar songs. Then we built two recommender systems--popularity-based and content-based--to recommend songs in the same cluster as the one an user inputs. Text-based interfaces were implemented to accept user input and present output.
Spotify Track Recommender System Project │ ├── README.md <- Project Description and Summary ├── spotify_tracks.csv <- Project Dataset ├── MSiA421_Final_Project.ipynb <- Project Jupyter Notebook └── MSiA421_Final Project.pptx <- Project Slides