Scientists publish together, companies enter into participatory networks and people connect via social media: millions of data points are generated daily, which can give us insights into interpersonal interactions and social processes.
There are two approaches to studying human behavior on the basis of this data: the first approach is primarily interested in the individual and individual phenomena and takes place conceptually mainly at the micro level. The other approach aims at understanding the effects of individual actions on the macro level of social coexistence. From the different perspectives, new methodological and algorithmic challenges arise for the operationalization of the questions, for the modeling of the data and finally for the analysis of the data.
In essence, we focus on two different objects of investigation: on the one hand, this professorship is dedicated to the structure of the above-mentioned social interdependencies and, on the other hand, we investigate the content of these interdependencies on the level of texts. Our methodological focus is on social network analysis, text mining and natural language processing (NLP). We apply these methods and also develop new algorithms in these areas.
Our special focus is the linking of structure and content in analysis. In order to better understand the behavior of individual actors in complex structures, a holistic perspective on the contexts in which the actors move is necessary. Situations have to be described comprehensively in order to enable novel insights. In addition to descriptive metadata such as time, place or activity information, this also includes the actual content of the structural linkages - e.g. texts in which people express their interests and views. Only when structural and contextual information are combined in their entirety will novel qualitative investigations be possible, from which colleagues in the humanities and social sciences can also benefit.
Network Embedding for Economic Issues
Fluid Ontologies of Contestation