Portrait of Benjamin Guinaudeau


Welcome to my Website! I am an Assistant Professor at Université Laval and a Faculty Research Affiliate with NYU’s Center for Social Media and Politics (CSMaP). Previously, I was a postdoctoral fellow at CSMaP, a PhD Student in Decision Science at the University of Constance (Germany) and a Research Scientist Intern at Meta Core Data Science in London (UK).

My research examines how emerging technologies reshape democratic representation in the 21st century! My work lies at the intersection of comparative politics, political communication, and data science. Starting from the premise that social media has transformed how political capital is created and distributed, my program advances three interlocking themes: modeling contemporary representation under social-media logics; auditing and explaining the biases that shape the production, demand, and distribution of political content on AI-driven, recommender systems; and building scalable methods and data infrastructure for social-media research even as platforms restrict data access. My recent work centers on TikTok, with earlier projects on Twitter (now X) and Meta platforms.

Beyond research, I teach both substantive and methodological courses in French, German, and English. During the 2022 French presidential election, I led the development of Poliverse, a platform designed to bridge academic knowledge and citizens through concise analytical briefs and interactive dashboards.

I am committed to fostering inclusion and diversity in academia. I strongly believe that domination and discrimination structures span everywhere in the academic world and that we can and should do better! In that sense, I have collected and analyzed conference data to document enduring disparities in political science, and I intend to expand this work in the future.

Outside my academic work, I am a compulsive #rstats programmer, a below-average pianist, and a dedicated foody! I also cherish any time I can spend in the Canadian wilderness sharpening my lumberjack skills, and I can’t get enough free space, wood chopping, and maple syrup.

Learn more about my research, teaching, software and CV!

Research agenda


Representation and Political Equality


My dissertation asks a simple democratic question: when party leaders keep their teams tightly unified, whose voices still get heard and whose are pushed to the margins? I show that cohesion inside parties is not just about ideology but about “political capital,” meaning the visibility, networks, and leverage that some representatives have and others lack. Leaders trade valuable perks and positions for loyalty, and that bargain can tilt policy toward those who already hold more capital, quietly shaping which voters are truly represented. Using clear evidence from Germany and the United Kingdom, I demonstrate when local constituencies win out, when party discipline prevails, and why the healthiest democracies find a balance where unity does not silence diversity. The normative takeaway is straightforward: for representation to be fair, influence inside parties should mirror citizens’ support, not celebrity or insider access, and leadership strength should be strong enough to govern without drowning out plural voices.

Technology and Politics


Democracy lives or dies by how attention, credibility, and accountability are distributed, and new technologies are reshaping that balance at scale. My research highlights the role of political capital, defined as the visibility, trust, and access that allow some voices to travel farther than others. Feeds and models now mint and move this capital, changing who sets agendas and who converts attention into influence. In that sense, TikTok represents the latest evolutionary stage of social media platforms: a video-first feed structured entirely by an opaque recommender system. My research on TikTok shows that it produces an attention oligopoly, in which a small group of creators captures an overwhelming share of attention. These actors are selected by performance rather than public mandate, forming a new elite that can frame issues ahead of parties and newsrooms. Moreover, centralized information flows deepen the risk of manipulation by state and corporate actors. These forces threaten representative democracy and urgently call for institutional redesign and technological regulation to ensure transparency, auditability, plural exposure, genuine platform competition, and independent research access.

Methodological Tools


I develop computational methods and tools for acquiring and analyzing political data, with core strengths in data engineering and text-as-data. My work spans statistical modeling, survey research, causal inference, web scraping, and applied machine learning in the social sciences. A central contribution of my dissertation is a measurement toolkit for democratic representation: an adaptive expert survey using pairwise comparisons with a Bayesian model to estimate legislators’ ideology in large chambers; multilevel regression and post-stratification to recover district-level preferences; and a reproducible proxy for individual political capital from pre-election Wikipedia traffic, complemented by a unified index of parliamentary appointments to trace how parties allocate influence. I build these pipelines mainly in R and Python and use JavaScript, C++, and Bash when needed. I also serve as the lead developer for NYU’s Center for Social Media and Politics (CSMaP) TikTok research effort, where I designed and deployed a method to collect a global random sample of TikTok content, now powering eight projects with collaborators at top institutions across Europe and the United States.