The case of music listening
Robin Cura, member of Géographie-cités, presented at the international conference CHI’22, a tool for the visual analysis of music listening history data, AMPLI, developed within the RECORDS project.
CHI’22, an international reference conference in the field of human-computer interaction and data visualization, organized by the ACM (Association for Computing Machinery), was held this year in hybrid mode, in person (New Orleans, LA) and by videoconference, from April 30 to May 5.
The corresponding paper, including many figures and videos, has been published within the extended abstracts of the conference:
Robin Cura, Amélie Beaumont, Jean-Samuel Beuscart, Samuel Coavoux, Noé Latreille de Fozières, Brenda Le Bigot, Yann Renisio, Manuel Moussallam, and Thomas Louail. 2022. Uplifting Interviews in Social Science with Individual Data Visualization: the case of Music Listening. In CHI Conference on Human Factors in Computing Systems Extended Abstracts
(CHI EA ’22
). Association for Computing Machinery, New York, NY, USA, Article 23, 1–9. https://doi.org/10.1145/3491101.3503553
See Robin Cura’s presentation: Uplifting Interviews in Social Science with Individual Data Visualization: the case of Music Listening. CHI EA ’22: April 2022 [VIDEO]
At this conference, Robin Cura presented a paper co-authored with other RECORDS project participants, on AMPLI (Amplifying Music listening Practices studies with Logs-augmented Interviews) software. Designed in a collaborative way between geographers, sociologists and computer scientists, AMPLI was tested during individual interviews on the music listening practices and musical tastes of about thirty consenting users of a streaming platform in France.
It helped the interviewers to prepare and conduct individual semi-structured interviews on various aspects of the listening experience, by allowing to visualize jointly (1) the users’ answers to a self-administered questionnaire survey and (2) their digital listening traces, resulting from the extraction of their complete listening histories on the Deezer streaming platform.
Robin Cura was the project manager of this experiment, as part of his post-doctorate (2020-2021) dedicated to this project, carried out within the UMR Géographie-cités.
An example of AMPLI’s exploration part. Here, there is both a graphical selection brushed in France Nor-mandy region, and selecting only the year 2020 of the data. We can see that the streams originating from this area comprised more Jazz music than usual for this interviewee (as illustrated by the “Tags” bar-plot in the bottom-right). The song the interviewee played the most in this area and during that time-frame was “Smile” by Jimmy Durante (as seen on the music player widget on the left menu).
By presenting interactive visual summaries of Deezer users’ data, AMPLI allows interviewers to immerse themselves in the music listening practices of their interviewees and thus open up avenues of discussion (elicitation) that make it possible to question the discrepancies between the respondents’ discourses on the subject of their cultural practices, and their actual practices as observed within the music streaming platform.
Beyond music listening practices, such a case study also and above all opens the way to a renewal of survey techniques through the use of “augmented” interviews on ordinary practices, taking advantage of the digital traces disseminated daily by everyone. In addition to the statistical analyses that can be made of the big data collected by the platforms, tools such as AMPLI allow to exploit the potential of these new sources of information in so-called “mixed” research, combining quantitative and qualitative methods.
Download R. Cura, A. Beaumont, J-S Beuscart et al. Improving social science interviews through individual data visualization: the case of music listening. CHI EA ’22: April 2022 , Article 23, p. 1-9, DOI: 10.1145/3491101.3503553
Collecting accurate and fine-grain information about the music people like, dislike and actually listen to has long been a challenge for sociologists. As millions of people now use online music streaming services, research can build upon the individual listening history data that are collected by these platforms. Individual interviews, in particular, can benefit from such data, by allowing the interviewers to immerse themselves in the musical universe of consenting respondents, and thus ask them contextualized questions and get more precise answers. Designing a visual exploration tool allowing such an immersion is however difficult, because of the volume and heterogeneity of the listening data, the unequal “visual literacy” of the prospective users, or the interviewers’ potential lack of knowledge of the music listened to by the respondents. In this case study we discuss the design and evaluation of such a tool. Designed with social scientists, its purpose is to help them in preparing and conducting semi-structured interviews that address various aspects of the listening experience. It was evaluated during thirty interviews with consenting users of a streaming platform in France.
More information on AMPLI
RECORDS (pRatiques dEs PubliCs des platefORmes De Streaming)
Through the RECORDS project (2020-2024) we aim to better understand the diversity of individuals musical taste, listening practices and effective music consumptions on streaming platforms. We develop indicators measuring the effects of recommendation (human and machine) on the listening history of users. To fulfill these objectives we combine traditional survey methods with big data analysis. The program is based upon a partnership between social scientists, computer scientists and one of the major players in music distribution in France, Deezer.
Read also: Influence of recommendation on music content consumption