Descriptif
We live in an era of digital traces. From our daily interactions with social media and large language models to the mass digitization of historical archives, human behavior now leaves an unprecedented digital footprint.
This explosion of data opens new frontiers for social scientists. Can we use large language models to classify—or even generate—social data? Do anti-immigration laws alter how people post online? How does war reshape world leaders’ rhetoric? Can simulations reveal whether network structures exacerbate or reduce gender segregation in labor markets? Ultimately, these digital traces allow us to ask entirely new questions and shed new light on classic social science questions.
This intersection is the foundation of Computational Social Science (CSS): combining social science theory, frameworks for handling massive and unstructured data, and advanced analytical methods to answer theory-driven questions.
This course is a broad, master’s-level introduction to CSS. Prioritizing breadth over depth, the goal is to help you develop a field-level “cognitive map” of the discipline. While not strictly a methods class, the curriculum is organized around six core methodological areas:
1.Social network analysis
2.Natural language processing
3.Machine learning
4.Generative AI
5.Online experiments and mass collaboration
6.Agent-based modeling
Each module introduces the technical foundations alongside the theories and substantive domains where these methods are frequently applied. Each module covers two weeks, and each week consists of a (brief) overview lecture, a discussion of an applied paper, and a hands-on application. If you are a CSS M2 student, the applications will prepare you to develop deeper technical expertise in your Semester 2 required and elective courses. If you are in a different degree program (e.g., QSD or Ingénieur), the applications will at least give you the foundation to do more systematic self-learning on your favorite CSS topics in the future.
All hands-on applications are done in the R statistical computing environment.
