Skip to main content

TagFlip: Active mobile music discovery with social tags

Resource type
Thesis type
(Thesis) Ph.D.
Date created
Recently, there has been a substantial surge in the availability of digital music through online subscription services. While these have given the user access to an almost endless variety of content, the sheer size of this choice space can also turn the user's selection process into a burden. While the purely algorithmic approach of recommending items with minimal user control has been the most popular method in commercial systems, several studies have shown that it can suffer from issues such as lack of transparency, limited user control, and pigeonholing the users. On the academic front, novel interfaces have been proposed that strive to alleviate the above issues. However, most such tools target large screens and often depend on complex user interactions. This is in discrepancy with the usual circumstances in which we listen to music, such as during commuting and work, and increasingly on our phones. We report on the design and evaluation of TagFlip. A mobile app featuring a novel interaction framework designed to fit within typical music listening scenarios, organically transforming them into more interactive music discovery experiences. Acknowledging the fleeting nature of music listening in the age of streaming and the smartphone, TagFlip was designed to require little effort from the users, while still granting them a high degree of control over the recommended music. This is done by describing each played song to the user through its most popular social tags and allowing the user to easily specify which of the tags should be considered for upcoming songs. To understand the merits of tag-based music discovery and especially our unique design, we compared TagFlip to Spotify's mobile app, both in a controlled lab experiment, and in the field. In these evaluations, TagFlip came out on top in both subjective user experience (control, transparency, and interaction) and our objective measure of number of interactions per liked song. Our field study participants found it fun and engaging to discover new tags and styles of music they had not heard of or liked before, and our logs indicated that they ended up actively seeking music within these new styles for significant portions of their time with the app. Our users found TagFlip to be an important complementary experience to that of Spotify, enabling more active and directed music discovery with minimal effort. Furthermore, we found that current mainstream approaches to music discovery may be incapable of exploiting the full potential of massive online music libraries.
Copyright statement
Copyright is held by the author.
This thesis may be printed or downloaded for non-commercial research and scholarly purposes.
Scholarly level
Supervisor or Senior Supervisor
Thesis advisor: Zhang, Hao
Member of collection
Download file Size
etd20980.pdf 14.14 MB

Views & downloads - as of June 2023

Views: 0
Downloads: 0