Context-aware Music Recommendation

Context-aware Music Recommendation

At the moment we are facing a fundamental change in the way people consume music: More and more people switch from private, mostly limited music collections to public music streaming collections containing several millions of tracks generating tons of data. The usability of such streaming services heavily relies on good recommender systems assisting users in discovering music they like. This makes the field of music recommendation and music information retrieval in a highly interesting topic for academia as well as industry. The DBIS Team focuses on context-aware music recommendation, exploiting data sources such as Twitter, last.fm. or Spotify. Our research is concerned with two types of context: Firstly, we focus on the current activity of a user while listening to music. Secondly, we are concerned with the cultural embedding of a user.

In our research, we analyze music listening behavior using machine learning techniques. The generated insights are integrated into music recommender systems, aiming at improving their prediction accuracy.

 

Public Datasets

For our research, we employ a variety of datasets that we have curated and utilized in our research and publications. We are happy to share the following datasets:

  • #nowplaying is a dataset which leverages Twitter for the creation of a diverse and constantly updated data set describing the music listening behavior of users. Twitter is frequently facilitated to post which music the respective user is currently listening to. From such tweets, we extract track and artist information and further metadata. You can find the dataset on Zenodo: https://doi.org/10.5281/zenodo.2594482 (CC BY 4.0).
  • The #nowplaying-RS dataset features context- and content features of listening events. It contains 11.6 million music listening events of 139K users and 346K tracks collected from Twitter. The dataset comes with a rich set of item content features and user context features, as well as timestamps of the listening events. Moreover, some of the user context features imply the cultural origin of the users, and some others—like hashtags—give clues to the emotional state of a user underlying a listening event. You can find the dataset on Zenodo: https://doi.org/10.5281/zenodo.2594537 (CC BY 4.0).
  • The Spotify playlists dataset is based on the subset of users in the #nowplaying dataset who publish their #nowplaying tweets via Spotify. In principle, the dataset holds users, their playlists and the tracks contained in these playlists. You can find the dataset on Zenodo: https://doi.org/10.5281/zenodo.2594556 (CC BY 4.0).
  • The Hit Song Prediction dataset features high- and low-level audio descriptors of the songs contained in the Million Song Dataset (extracted via Essentia) for content-based hit song prediction tasks. You can find the dataset on Zenodo: https://doi.org/10.5281/zenodo.3258042 (CC BY 4.0).

Team

Publications

2012

Bib Link Download

Eva Zangerle, Wolfgang Gassler and Günther Specht: Exploiting Twitter's Collective Knowledge for Music Recommendations. In Proceedings of the 2nd Workshop on Making Sense of Microposts (#MSM2012): Big things come in small packages, Lyon, France, 16 April 2012 (in connection with the 21st International Conference on World Wide Web), pages 14-17. 2012