Recommender systems, like the one Spotify uses, tend to be opaque and don’t show users how recommendations are calculated or how they relate to the rest of the items in a dataset.
To tackle this, I present the MusicExplorationApp. This web application uses a map metaphor to show users where their preferences and recommendations are located on a map of music. In addition, this map can be used to explore a space of about 9 000 artists.
To read more about the MusicExplorationApp and a user study we conducted see:
Johannes Kunkel and Jürgen Ziegler. 2023. A comparative study of item space visualizations for recommender systems. International Journal of Human-Computer Studies 172, C (Apr 2023). https://doi.org/10.1016/j.ijhcs.2022.102987
The supplementary material of the application and the experiment conducted can be found at:
Kunkel, Johannes; Ziegler, Jürgen (2022), “Supplementary Material for: “A Comparative Study of Item Space Visualizations for Recommender Systems””, Mendeley Data, V2, http://dx.doi.org/10.17632/hw39d5kxs6.2
This project is licensed under the terms of the MIT license.