Spotify You (Group 49)
Table of contents
Overview
For our final project, we created a best-performing model to help classify songs for Grace’s playlist. We used the Spotify API to download features for songs Grace selected as playlist-worthy and playlist-unworthy. After getting a grasp on the data, we built classifiers to predict whether a song should be in or out of her playlist. The Decision Tree Classifier with Boosting was the best-performing model, improving our accuracy in the test set from 51.6% to 93.0%. We used this model to classify a fresh set of songs for Grace, and she loved listening to her augmented playlist!
Website Navigation
Our motivations for the project as well as our literature review can be found on the background page. A description of the data as well as our exploratory data analysis can be found on the data exploration page. Our models can be found on the models page. Our results, analysis, conclusions, and suggestions for moving forward can be found on the conclusions page.
About Us
We are Tejal Patwardhan, Akshitha Ramachandran, and Grace Zhang, Group 49 for CS109A. Special thanks to Pavlos Protopapas, Kevin Rader, and Rashmi Banthia for their assistance.
