New Endpoints: Audio Features, Recommendations and User Taste
Posted March 29th, 2016
Over the two years since Spotify acquired The Echo Nest, we have folded much of the power of The Echo Nest into Spotify’s stack. With today’s announcement of the wind down of The Echo Nest API, we’re happy to launch three new endpoints to bring some of the best parts of The Echo Nest’s API to Spotify.
Spotify runs a suite of audio analysis algorithms on every track in our catalog. These extract about a dozen high-level acoustic attributes from the audio. Some of these are well-known musical features, like tempo and key. Others are more specialized, like speechiness or danceability. Our new Audio Features API makes all of this analysis available for every track in Spotify’s catalog.
Our first recommendations API generates listening experiences based on a set of seed artists, tracks, genres, or a combination of all three. If you would like to tweak your results further, you can filter on audio features and track popularity to easily build a great set of tracks for your application.
User Top Artists and Tracks
We are introducing a new API endpoint that will allow developers to retrieve the top artists and tracks for a Spotify user once the user has granted access. This API will make it easier for developers to create personalized experiences for their users.
This data is calculated based on the listening and other behavior of users and we’re computing it over three different time ranges. The short-term dataset represents only the most recent actions of a user. The medium-term dataset reflects a broader, slower changing sense of their listening preferences. The long-term dataset incorporates the user’s entire listening history.
Take a look at the docs.