CSS Team Winter Symposium

At first sight, the debate surrounding #covid19 polarized society into two camps: on the one hand, the majority of citizens who supported the political crisis management, and on the other, so-called “corona activists” who protested against protective measures. The latter appeared to be influenced by a massive spread of fake news and conspiracy ideas, particularly on the internet. Less known are the concrete social practices of news reception in digital arenas such as Twitter, which may have influenced the ways in which users of the platform connected or antagonized each other through sharing and commenting on tweets. This study takes a close look at the discursive construction of corona narratives in response to Covid-related news in France. The focus is on the semantic framing in quotes shared as tweets with a comment. When many users share and comment on the same message, “quote trees” emerge, a communication phenomenon similar to the “broken telephone game”, where a message changes its meaning and tone as it passes from one user to another. By studying this phenomenon on different topics along 2020, we aim to show that punctual controversies have emerged in a context of overlapping debates and in response to President Macron’s televised citizens’ speeches, which may have influenced the emergence and evolution of political counter-speech. We find successive phases of debate in which Twitter users gradually turn their backs on the established representatives of government, the media and science. It seems as if local institutions, alternative experts and self-produced media content are mobilized to regain orientation and agency.

When analysing social networks, it is often helpful to consider attributes that are known about the actors, i.e. metadata, to provide further insights into the underlying network dynamics. For example, when considering online public discourse represented by an interaction network of social media users we might not only be interested in the links between the user nodes, but also in additional user information, such as demographics or political preference, and the interplay between such user categories and link formation. One way in which we can integrate metadata in network analysis is to explore the relationship between metadata and what we call the network’s block structure — a division of the network into groups of nodes that are similar in terms of how they are connected to each other and the rest of the network. In the network science literature, the prevalent assumption of an intrinsic connection between metadata and block structure has faced scrutiny in recent years. Nevertheless, a tool for measuring the metadata-structure relationship for the purpose of systematic comparative meta analyses has been notably lacking. In this talk, I present a novel approach to measure the relationship between metadata and network structure. Our tool utilises Stochastic blockmodels and concepts from information theory to quantify two things: (i) the strength of the connection between node metadata and the block structure of networks, and (ii) the likely structural arrangement of metadata partitions. In the talk, I introduce the method and demonstrate it on some social networks, including political discussion networks from Twitter.

In this talk, I’ll present the progress of my first two months of work in the CSS team. I’ll present a first draft of a visualization device designed to represent dynamic stochastic block models. I’ll come back to the conceptual and ontological choices behind this draft. In particular, I’ll focus on the development of a visual stabilization algorithm for networks of blocks evolving over time. This work is aimed at better understanding the temporal evolution of various empirical socio-semantic systems.

This research explores how users of the French music streaming platform, Deezer adopt different platform affordances such as personal playlists or rather algorithmic or editorial playlist, to both explore and exploit music on-platform. Leveraging a large-scale 500K user set, we commence from a comprehensive analysis of socio-demographic variations with respect to users’ affordance usage on-platform. Our work then progresses to present an exploration of how firstly, low level behavioural dynamics such as activity and redundancy and then later, concentration dynamics (both semantic and item centric) correspond to different affordance adoption strategies thus, grounding our work within the novel strand of research appraising user diversity profiles within multi-affordance environments.

My study utilizes a database encoding the music consumption of 50,000 users on the Deezer platform over three years (2020-2021-2022). The primary goal of my study is to uncover insights into the spatial configuration of the music industry and its impact on commercial success. While contemporary research leveraging this kind of digital traces is often hailed as groundbreaking, the discipline of geography has encountered various debates surrounding the increasing mathematization of its methodologies. Traditionally, human geography, as a scientific field, has embraced a deeply idiographic approach grounded in fieldwork and observation. Consequently, human geography has gradually divided into two methodologically opposed fields: qualitative, or classical, geography and quantitative geography, commonly referred to as spatial analysis. My study naturally falls within the latter category. Building on prior research in the geography of rap music, which underscores the role of local connections in an artist’s commercial success, my study empirically explores the commercial success associated with declared strategies of local musical collaborations. It delves into the role of geographical proximity between artists in the intersection of social cohesion and commercial success. The findings contribute to formulating precise and relevant research questions, guiding forthcoming interviews with key stakeholders in the music industry.

Our work is based on the premise that the music recommendation activity should no longer be analysed in exclusively cultural terms, at the risk of overlooking its organisational issues. This change of perspective makes it possible to reconsider the way in which certain values are instantiated in recommendations, especially when they are algorithmic. In a study based on a dozen interviews with Deezer employees, we show, for example, that exploration and diversification are not aims in themselves – even if they are high ideals – but are entirely appropriate means for recommending music and, by extension, getting people to consume it. This commercial objective is the driver of competition between recommendation modules, as well as collaboration between the actors involved in their design. Our main result highlights the importance of algorithms in organising both music recommendation (competition) and organisation (collaboration).

Since the 1990s, the Environmental Kuznets Curve (EKC) hypothesis posits an inverted U-shaped relationship between pollutants and economic development. The hypothesis has attracted a lot of research. We provide here a review of more than 2000 articles that have been published on the EKC. We aim at mapping the development of this specialized research, both in term of actors and of content, and to trace the transformation it has undergone from its beginning to the present. To that end, we combine traditional bibliometric analysis and semantic analysis with a novel method, that enables us to recover the type of pollutants that are studied and the empirical claims made on EKC (whether the hypothesis is invalidated or not). We principally exhibit the existence of a few epistemic communities that are related to distinct time periods, topics and, to some extent, proportion of positive results on EKC.