Detailed information about the course
| Title | Winter School on Computational Social Science |
| Dates | 26-30 January 2026 |
| Lang |
Workshop language is English |
| Responsible | Mathias HUMBERT |
| Organizer(s) | Pr Mathias Humbert (UNIL) Pr Yash Raj Shrestha (UNIL) Pr Pascal Felber (UniNE) Pr Valerio Schiavoni (UniNE) |
| Speakers |
|
| Description | This winter school will addresse the topic of computational social science with a specific focus on the web, social media, and politics. Prof. Alexandre Bovet (University of Zurich) will present a network science and complex systems perspective to understand how platforms shape political discourse. In the hands-on session, we will then explore different methods to quantify polarization, using LLMs to further analyze stance and animosity. Prof. Juhi Kulshrestha (Aalto University) will explain how we can combine diverse data sources such as passively collected browsing traces, online questionnaires, and experiments to study our Internet-mediated lives. In the hands-session, we will analyze a dataset of browsing traces with their sociodemographic information, political beliefs, and mental health states to examine various associations between Internet use and attittudes, beliefs, and mental health. Prof. Pedro Ramaciotti Morales (Sciences Po Paris) will present a computational framework for analyzing the political impact of online platforms and algorithms through systematic political characterization of digital content and actors across countries. Finally, Prof. Ingmar Weber (Saarland University) will present two alternative methods for collecting social media data in the "Post-API" age, namely through advertising platforms or personal data donations.
|
| Program | Monday 26.01
Tuesday 27.01
Wednesday 28.01
Thursday 29.01
Friday 30.01
Session A, Prof. Alexandre BovetPolarization on social media, from network modeling to real-world analysisSocial media platforms have transformed civic dialogue, becoming vital spaces for political exchange and democratic participation. While initially heralded as democratizing forces that amplify marginalized voices and enhance civic engagement, these platforms now reveal a more complex reality. Recent evidence suggests that they simultaneously enable democratic participation and pose serious challenges, fueling political polarization, accelerating the spread of misinformation, and eroding civil discourse. These impacts vary significantly across different platforms and sociocultural contexts. Understanding precisely how platforms shape democratic discourse is a challenging task. Observational and experimental studies often yield conflicting results. Here, I will take a network science and complex systems perspective to elucidate how these seemingly different results can be understood and what can be done better to comprehend the role of social media on society. In the practical session, we will explore methods to quantify several polarization measures, using LLMs to analyze stance and animosity, and network-based measures of interventional polarization. We will also examine how these different aspects of polarization help us better understand the whole system. Session B, Prof. Juhi KulshresthaExploring How Online Behaviour Shapes Opinions and Wellbeing using mixed-methodsIn today's digital world, the internet and web-based technologies shape nearly every aspect of daily life. This session explores how we can combine diverse data sources such as passively collected browsing traces, online questionnaires, and experiments to study our internet-mediated lives. Using mixed-methods approaches, we will learn how to examine human behaviour online and its impact on individuals' opinions, behaviours, and wellbeing both online and offline. In the hands-on session led by Yajing Wang, we will provide a dataset of individuals' web browsing traces combined with their sociodemographic information, political beliefs, and mental health states collected via surveys. We will analyze the data to examine the associations between individuals' internet use and their attitudes, beliefs, and mental health. Participants will also have the opportunity to analyze their own browsing history to infer some of these associations for themselves. Session C, Prof. Pedro Ramaciotti MoralesReading world politics through online data traces and applications to social platform and AI regulationThese lectures will present a computational framework for analyzing the political impact of online platforms and algorithms through systematic political characterization of digital content and actors across countries. We will develop multidimensional measurement approaches that map political positions across ideological spectra (e.g., Left-Right) and issue-specific dimensions (e.g., immigration, environmental policy), enabling cross-national comparative analysis and temporal. We will apply this framework to investigate key issues in digital politics, including misinformation dynamics, algorithmic bias in content recommendations, and quantitative polarization measurement. We will also explore how online behavioral data can serve as a proxy for understanding offline political structures, particularly regarding multidimensional polarization patterns and the evolving dimensionality of political spaces. This dual approach will provide generalizable computational tools for the study of online politics in computational social sciences, applicable to a diversity of settings, while exploring key results pertaining to platform and AI regulation specifically, and to digital compared politics more generally. Session D, Prof. Ingmar WeberAdvertising and Data Donations: Alternative Methods for Collecting Social Media DataCollecting large amounts of social media data is getting increasingly difficult, leading researchers to calling this the "Post-API Age". In our two sessions, we'll show two not-so-obvious ways of collecting social media data despite these limitations. First, we'll show how basic user counts can be obtained through the advertising platforms of Meta, TikTok, LinkedIn and even Twitter/X. While such data does not contain any actual posts, it is often still useful for various modelling tasks, e.g. related to tracking digital gender gaps, monitoring migration, or mapping wealth inequalities. Second, we'll demonstrate how study participants can be recruited and incentivized to "donate" their personal social media data. Such data collections are particularly powerful when combined with surveys that help, e.g., assess whether a user sees their own social media use as addictive or not. Furthermore, advances in multi-modal foundation models have made the analysis of multimodal content increasingly accessible. The first session will be more lecture like and given by Ingmar Weber. The second session will be hands-on and given by Brahmani Nutakki. |
| Location |
Hôtel Suisse, Champéry |
| Map | |
| Information | |
| Registration | Deadline for registration: 15.12.2025 A mandatory fee of 100 CHF per CUSO students is required and to be paid before the beginning of the Winter School. Please note that, if you register online but do not pay the 100 CHF and do not come to the winter school, you will also have to pay the lodging and meals costs.
|
| Places | 35 |
| Deadline for registration | 15.12.2025 |
| Contact |
short URL
URL onepage