Spotify Song Recommender
Program project, build a recommender system to recommend similar songs using Spotify's API.
Published on November 10, 2019 by Claudia Chajon
recommender Spotify machine learning
0 min READ
–
Recommendation model using Spotify songs and features
– —
See the work
Implemented a K-nearest neighbors model to recommend songs to users based on their profiles. The features that make up a song are unique to Spotify.
Songs were recommended based on how alike they were based on selected features.
Spotify engineers their own features and while they might have funny names, the features are great metrics to compare songs to one another.
Features includes:
- danceability
- energy
- acousticness
- instrumentalness
- liveness
- loudness
- speechiness
- valence
These four radar charts show 4 different songs that the KNN model found to be similar in features.
