Computational Social Science (CSS) is a new area of research which aims to analyze socio-cultural phenomena through new kinds of data and novel technologies. The formation of the CSS department at GESIS in 2013 can be regarded as part of and reaction to the digital transformation of society, in the course of which transactions in the digital world and structures of the physical world become ever more closely intertwined. Digital behavioral data reflect these processes. This kind of data incorporates valuable information and, at the same time, poses various methodological challenges to the social sciences: the data is (unprecedentedly) vast, often unstructured, dynamic, and of high resolution; its biases are as unknown as its true potential for explaining social structures is yet unexplored.
The overall goal of the department is to develop and provide methods and tools for the social sciences to mine digital behavioral data (especially from the web), to collect sensor data (e.g., via RFID chips or cell phone apps), and to combine this new kind of data with traditional survey data in order to improve the analyses of a broad range of socio-cultural phenomena. The CSS department tackles these tasks with approaches from machine learning, text and data mining, and network analysis.
Current research topics are political communication and election campaigning, the emergence of social and cultural inequalities and biases, and collaborative forms of knowledge production. The development and application of novel methods generate new insights and best practices and they also prepare the ground for research based infrastructure services.
GESIS promotes the formation and networking of an interdisciplinary CSS community through establishing transparency of and providing open access to research processes (Open Science), by organizing international conferences and symposia, and by offering training courses in data science and CSS methods.
The CSS department consists of the following teams:
The task of the Knowledge Discovery team is the acquisition, description and enrichment of digital behavioral data that help address research questions from the social sciences. To this end, the team develops and applies scalable methods from the areas of machine learning, data and text mining, and semantic web technologies. Both the produced data and the methods developed to acquire, describe and enrich this data are made available for further research and service purposes.
The mission of the Data Science team lies in the development and evaluation of models for data analysis that are based on digital behavioral data. The team mainly focuses on generative network models for explaining and predicting behavior of subpopulations (e.g., interactions between scientists of different gender or discipline) as well as on elaborating statistical models for sequential human behavior (e.g., the decisions made when navigating on the web or individual movement in urban surroundings).
The Social Analytics and Services team focuses on the empirical analysis of social behavior, especially in online communities. To this end, the team develops techniques to enhance our understanding of phenomena such as political communication, collaborative content generation on Wikipedia, or the dissemination of scientific knowledge into society. Moreover, the team develops innovative data-driven tools and services and organizes conferences and networking events for the social sciences.