On Skipping Behaviour Types in Music Streaming Sessions

Abstract

The ability to skip songs is a core feature in modern online streaming services. Its introduction has led to a new music listening paradigm and has changed the way users interact with the underlying services. Thus, understanding their skipping activity during listening sessions has acquired considerable importance. This is because such implicit feedback signal can be considered a measure of users' satisfaction (dissatisfaction or lack of interest), affecting their engagement with the platforms. Prior work has mainly focused on analysing the skipping activity at an individual song level. In this work, we investigate different behaviours during entire listening sessions with regards to the users' session-based skipping activity. To this end, we propose a data transformation and clustering-based approach to identify and categorise skipping types. Experimental results on the real-world music streaming dataset (Spotify) indicate four main types of session skipping behaviour. A subsequent analysis of short, medium, and long listening sessions demonstrate that these session skipping types are consistent across sessions of varying length. Furthermore, we discuss their distributional differences under various listening context information, i.e. day types (i.e. weekday and weekend), times of the day, and playlist types. CCS CONCEPTS• Information systems → Recommender systems.

Podify: A Podcast Streaming Platform with Automatic Logging of User Behaviour for Academic Research

“…They should offer academic and independent researchers the capability to perform user studies and advance the field. However, there is currently no podcast streaming platform with a content-based podcast search engine [4] for academic research and no data collection with logged user behaviour [15]. To address these limitations, in this paper, we present Podify.…”

“…The dataset's release was in conjunction with the TREC Podcast Track [11], which ran in 2020 and 2021, where two shared tasks were defined: retrieval of fixed two-minute segments and episodes summarisation. Although the dataset's great applicability to various tasks in fields such as speech and audio processing, natural language processing, information retrieval, and computational linguistics, it is unsuitable for those where logged user behaviour is required [15]. This is the case of analyses of user information needs, their characteristics and behaviour, relevance, search, recommendation, and personalisation systems.…”

“…This platform allows researchers to conduct user studies in the podcast domain to alleviate the lack of user interactions in the Spotify Podcast Dataset [15]. This work can foster academic research toward podcast consumption patterns and listening activities and thus elicit research in novel personalisation techniques.…”