Spotify Million Playlist Challenge

Team NPK


Karina Huang, Nicholas Stern, Phoebe Wong

Harvard AC209a Final Project Fall 2018


This GitHub Page serves as a demo page for our final project. The complete record of analysis is hosted on our github repo.

Refer to the following description for navigation:

Overview

A description of the Spotify playlist recommendation problem, touching on the main difficulties and the primary motivations for predicting song preferences on music platforms.

Data & Processing

A closer look at the structure of the data through some exploratory analysis, along with a description of how the data were cleaned and processed.

Modeling & Results

A breakdown of the models used for playlist recommendation, with detailed explanations behind each algorithm and its performance. The models we implemented are:

Discussion & Future Work

a conversation about what the next steps would be in exploring techniques for playlist recommendation, as well as extra ideas we had, but have yet to implement.