The lecture by Fabian Flöck from GESIS – Leibniz-Institute for the Social Sciences will take place at the Complexity Science Hub Vienna in the Salon.
If you are interested in participating, please email to office@csh.ac.at
Abstract
Using Social Media data to complement more “traditional” data sources, foremost survey data, has increasingly become of interest for social scientists in different fields. At GESIS – Leibniz Institute for the Social Sciences, we have been looking at the (conditions for the) useful applications of such a marriage of data collection instruments from Social & Computer Science standpoints.
In my talk, I will highlight chances and fallacies of Social Media data for answering research questions about a broader target population, particularly from the perspective of the total error framework in survey methodology. This includes representation and measurement errors, plus the new challenges arising with Social Media.
Secondly, I will outline ongoing research at the GESIS CSS department addressing some of these challenges. One is a project aiming at the “n=all” sample collection of relevant Twitter content regarding key election actors and topics; and connected to it our efforts providing such a collection as a service to Social Scientists. I will also cover a project in which we made advances in inferring Social Media users’ demographic attributes in different languages with a multimodal DL approach and using them in reweighing techniques for sample bias reduction.