| Dozent/in |
Prof. Dr. Petter Törnberg |
| Veranstaltungsart |
Workshop |
| Code |
HS261683 |
| Semester |
Herbstsemester 2026 |
| Durchführender Fachbereich |
Diverse |
| Studienstufe |
Master
Doktorat |
| Termin/e |
Fr, 30.10.2026, 09:15 - 16:30 Uhr, 3.B52
Sa, 31.10.2026, 09:15 - 16:30 Uhr, 3.B52 |
| Umfang |
Blockveranstaltung |
| Inhalt |
This two-day workshop introduces a computational social science approach to comparative political analysis, focusing on how large language models (LLMs) and natural language processing (NLP) can be used to generate data from political text. Participants will work with cross-national datasets of political communication, including social media posts from politicians and parties, and party manifesto data.
Through hands-on exercises in Python, students will use LLMs and NLP techniques to measure concepts such as misinformation, toxicity, and political framing at scale. These derived measures will then be combined with traditional comparative datasets (e.g., V-Dem, Parlgov, CHES) to enable cross-country and cross-party analysis. The workshop emphasizes how computational methods can transform unstructured text into structured variables suitable for comparative research.
|
| Lernziele |
1. Understand how LLMs and NLP can be used to generate quantitative data from text
2. Gain hands-on experience applying computational methods in Python
3. Learn how to construct comparable measures across countries and parties
4. Combine text-derived variables with traditional political science datasets
5. Design and implement small-scale computational comparative research projects
6. Critically assess the validity and limitations of LLM-based measurement
|
| Voraussetzungen |
Basic knowledge of social science and quantitative methods. Some familiarity with Python is recommended but not required. |
| Sprache |
Englisch |
| Begrenzung |
Only Master or doctoral students (postdoctoral researchers can sign up via lumacss@unilu.ch). Priority for LUMACSS students if there are too many registrations. |
| Anmeldung |
***Important*** To earn credits, you must register for the course via UniPortal. Registration is open from two weeks before to two weeks after the start of the semester. You can withdraw from the course after this period by notifying the lecturer and lumacss@unilu.ch. You can find the registration details here: http://www.unilu.ch/ksf/semesterdaten |
| Leistungsnachweis |
To obtain 1 ECTS, students must:
• Actively participate in all workshop sessions and hands-on exercises
• Complete a small-scale computational project using the methods introduced in the workshop |
| Abschlussform / Credits |
Bestätigte Teilnahme / 1 Credits
|
| Hörer-/innen |
Nein |
| Kontakt |
lumacss@unilu.ch |
| Material |
Materials will be shared by the lecturer via email in advance during the week of the workshop. |