Familiarizing with Music: Discovery Patterns for Different Music Discovery Needs

TL;DR

The paper investigates how users' self-reported interest in unfamiliar music ($IUM^u$) shapes discovery and exploration patterns using Deezer data paired with Gold-MSI survey responses. It introduces cluster-based tracking of tracks via $k$-means on SVD-derived embeddings of track co-occurrence, validated against editorial playlists, and defines newly discovered clusters after a six-month baseline followed by a one-week discovery window. Key findings show that higher $IUM^u$ correlates with more clusters discovered and greater initial diversity, while exploration dynamics reveal early tracks are more popular and less representative, with a trend toward more genre-representative tracks as exploration progresses; effects vary with $IUM^u$. These insights enable inference of unfamiliar-music interest from streaming data and inform recommender systems that guide exploration with attention to track popularity and genre representativeness. The work highlights practical implications for personalized music discovery and lays groundwork for future studies across broader discovery dimensions and time scales.

Abstract

Humans have the tendency to discover and explore. This natural tendency is reflected in data from streaming platforms as the amount of previously unknown content accessed by users. Additionally, in domains such as that of music streaming there is evidence that recommending novel content improves users' experience with the platform. Therefore, understanding users' discovery patterns, such as the amount to which and the way users access previously unknown content, is a topic of relevance for both the scientific community and the streaming industry, particularly the music one. Previous works studied how music consumption differs for users of different traits and looked at diversity, novelty, and consistency over time of users' music preferences. However, very little is known about how users discover and explore previously unknown music, and how this behavior differs for users of varying discovery needs. In this paper we bridge this gap by analyzing data from a survey answered by users of the major music streaming platform Deezer in combination with their streaming data. We first address questions regarding whether users who declare a higher interest in unfamiliar music listen to more diverse music, have more stable music preferences over time, and explore more music within a same time window, compared to those who declare a lower interest. We then investigate which type of music tracks users choose to listen to when they explore unfamiliar music, identifying clear patterns of popularity and genre representativeness that vary for users of different discovery needs. Our findings open up possibilities to infer users' interest in unfamiliar music from streaming data as well as possibilities to develop recommender systems that guide users in exploring music in a more natural way.

Figures (4)

• Figure 1: Histogram of $\text{IUM}^u$, i. e., of replies to the question "I'm intrigued by musical styles I'm not familiar with and want to find out more.". $\text{IUM}^u=1$: fully disagree; $\text{IUM}^u = 5$: fully agree.
• Figure 2: Average number $N^u$ of clusters explored in $[t_1, t2]$ (one week after initial six months) for different values of $\text{IUM}^u$. Central dots and error bars represent mean and standard errors estimated with bootstrapping (see Section \ref{['sec:methodology_previously_unexplored_music']}).
• Figure 3: Diversity and stability of music preferences for different declared interest in unfamiliar music $\text{IUM}^u$. Central dots and error bars represent the mean and standard errors estimated with bootstrapping.
• Figure 4: Clusters exploration displayed as a function of the time-rank $\text{ExpR}^u_m$ of tracks listened to in a newly discovered cluster. These plots are computed over all users.