![[OLD FALL 2020] 15-104 • Introduction to Computing for Creative Practice](https://courses.ideate.cmu.edu/15-104/f2020/wp-content/uploads/2021/08/cropped-notcurrentsemester.jpg){kind=small}
![[OLD FALL 2020] 15-104 • Introduction to Computing for Creative Practice](https://courses.ideate.cmu.edu/15-104/f2020/wp-content/uploads/2021/09/stop-banner.png)
Designed by Frederic Vavrille, Musicovery is an interactive music service program that personalize music recommendations to various types of listeners by algorithms.
Listeners can start by choosing a mood or a radio as the initial metrics. Then the sound navigate the listeners to navigate between different artists.
Many music service engines are only based on collaborative filtering and user’s context, facing problem such as creating clones of playlist with very few artists, repetitively shuffling highly popular songs and cold start issues. To allow listeners to wander strategically out of their song/artist preferences, Musicovery optimize the system with more customized metrics to measure diversity. By concentrating less on the tops of the playlist, increasing higher variety and disparity, the service provides not only users based playlist but also context based playlist with more different navigations on genre. It also measures metrics including skips to analyze listeners’ level of engagement to each music. This increase the accuracy of the listeners’ preferences significantly.
