Summer School Series on Methods for Computational Social Science
CSS Summer School
Leader: Prof. Dr. Claudia Wagner
Scientific unit: Computational Social Science (CSS)
Computational Social Science (CSS) is an interdisciplinary research field in which scholars of different communities work on similar topics. However, the lack of meeting opportunities and the lack of methodological understanding across disciplinary boundaries hinder many collaborations.
The project aims to address this issue by setting up an interdisciplinary summer school series for young scholars from different disciplines that work in the same thematic area. The training program will focus on methods for the three main types of data in CSS research (textual data, multimedia data, digital traces data) and the main goal is to equip junior scholars with the methodological knowledge that is required to approach current and future societal challenges.
Each summer school will focus on a specific type of data (behavioral trace data, text and multimedia), corresponding methods and a selected research area from the social sciences (social and cultural phenomena, bias, radicalization, polarization and conflicts).
In the first year, the “CSS methods for behavioral trace data” will focus on how to generate and analyze this type of data to gain insights into social phenomena such as mobility, the formation of consensus, behavioral changes, and the emergence of prosocial behavior. Digital trace data are produced by humans as a side product of their daily activities on the Web (e.g. click streams) or when using new technologies (e.g., fitness devices or sensors). However, researchers can also generate digital trace data by running massive online experiments or field experiments.
In the second year, the “CSS methods for text” summer school will teach methods to analyze large amounts of text (e.g., sentiment analysis, latent semantic models, causal inference methods for text data) and students will apply these methods to answer social science questions about conflicts, radicalization, polarization and bias. With the advent of the web and social media, the amount of text that is published online and is available to the general public, has increased enormously. Personal stories in blogs, Wikipedia articles and discussion pages, and online discussions on Twitter and in newspaper forums, provide a rich source of information that can be exploited to understand e.g. the emergence of conflicts, biases and radicalization online.
The “CSS methods for multimedia” summer school in the third year will go beyond text and will teach methods that can be used when analyzing multimedia data. Multimedia data such as images about food, memes, jokes, and videos, that are shared or become popular in different geographic areas, provide a valuable and highly underexploited source of data that allows to gain insights into cultures, subcultures and their dynamics over time.
Besides offering a unique opportunity for transferring interdisciplinary methodological knowledge, the summer school will also allow students to network and discuss their research or research proposals with experts in the field. To ensure that students get to know each other and receive feedback on their research, it will be organized a poster session at the first day where each participant will present a poster about his/her research or research plans. This will allow students to find other participants and experts with similar interests.
To foster collaborations between students, the project partners will initiate team projects during the course of the week. Lecturers will propose topics and mentor 1-2 teams. At the end of the week all students have to present the outcome of their project (e.g., results from an empirical analysis, a research design to answer an open question, comparison of different methods, ideas for new models and methods).
More information on: summerschool.computationalsocialscience.eu
- Dr. Nicola Perra, PH.D, University of Greenwich, International Business School and Economics, UK
- Dr. Emilio Ferrara, University of Southern California, Los Angeles, CA, USA
- Dr. Michael Macy, Cornell University, Ithaca, N.Y., USA