1. médialab Sciences Po
  2. News
  3. Joke Maths: A Proposed Method for the Modelling and Analysis of 91 Netflix Stand-up Comedy Specials – Time, Bodies, Laughter, and Themes in Contemporary American Stand-up

Joke Maths: A Proposed Method for the Modelling and Analysis of 91 Netflix Stand-up Comedy Specials – Time, Bodies, Laughter, and Themes in Contemporary American Stand-up

This seminar session will be presented by Yaelle Zribi, who will discuss her research on computational and critical analysis of stand-up comedy.

Event, Research Seminar

salle Goguel, 56 rue des saint pères, 75007 Paris

Abstract

This presentation introduces a methodological proposal for the computational and critical analysis of stand-up comedy, based on a corpus of 91 Netflix specials. I combine multimodal modelling (computer vision, audio signal analysis, and natural language processing) with exploratory ethnography conducted in New York, as well as more traditional approaches to performance analysis through close reading. The resulting pipeline detects recurrent body poses, identifies laughter segments, and models thematic content, allowing for a systematic approach to timing understood as the fine coordination of gesture, speech, and audience response. Beyond the technical contribution, the project engages with aesthetic and epistemological questions: how performance can be rendered computable, to what extent humour and play can (or cannot) be communicated to machines, and what challenges arise when artistic forms are approached through computational systems. This research opens a dialogue between theatre and performance studies, digital humanities, and machine learning, proposing both a tool and a critical reflection on the modelling of live performance: what biases are at play, what potential emerges, and what it means to see through, with, and by machines.

Biography

Yaelle Zribi studied economics, mathematics, and sociology before completing a bi-disciplinary degree in literature and history at Université Paris-Cité and Trinity College Dublin. She completed her graduate training in Digital Humanities at Université Paris, Sciences & Lettres, with dual training in computational methods (École des chartes–PSL) and art theory (ENS–PSL). Her research focuses on the visual analysis of large-scale corpora of artistic works, combining quantitative modelling with aesthetic perspectives.

Practical informations

The session will take place on Tuesday, September 23, 2025, from 2:00 PM to 4:00 PM, in person and in English at Sciences Po, Room Goguel, 56 rue des Saints-Pères, Paris 75007.

Registration is mandatory via this link.