Digital Behavioral Data

The digitization of society is continuously generating new kinds of data, tracing individual human behavior as well as social, societal and political processes in unprecedented detail. Combined with new methods from the fields of data science and machine learning, these digital behavioral data enable scientists to effectively study the digital society and its processes.

Specifically for the social sciences, digital behavioral data offer various new opportunities for analyzing social and political phenomena: from global networking to political opinion formation and polarization, to describing networks, group behavior and global cultural patterns. Digital behavioral data can also be used to explore the dynamics of socio-technical systems by investigating how web platforms and their algorithms reinforce existing societal tendencies or create their own social realities. These systems can significantly influence our interaction patterns online and offline, for example through search recommendations, (biased) rankings or communication cascades. Consequences can include the promotion of echo chambers, disinformation or hate comments, as well as the reinforcement of structural inequalities.

In 2013, GESIS established the Computational Social Science department as a foundation to exploit the potential of digital behavioral data for the social sciences. Since then, GESIS has been continuously working on new services for social scientists across all departments, providing basic training, consulting services, access to selected digital behavioral data and methods to analyze them.

Our long-term goal is a strategic, institute-wide expansion of GESIS' services in the area of digital behavioral data across all phases of the research data cycle. A special focus in this endeavor is on linking innovative digital behavioral data with established survey methods. This goal can only be achieved by taking into account the unique ethical, legal, technical and methodological challenges of new types of data. In cooperation with the scientific community and international partners, we are thus developing innovative infrastructures to make digital behavioral data a powerful tool for social science research.

Selected activities and services for digital behavioral data: 

  • GESIS notebooks: Open Science platform for browser-based analysis of digital behavioral data 

  • WikiWho: Tool for quantifying peer-produced knowledge on Wikipedia 

  • TweetsKB: Dataset with more than 2 billion preprocessed and annotated Tweets from 2013 onwards 

  • TweetsCOV19: Semantically preprocessed corpus of 8 million tweets about the COVID-19 pandemic 

  • ClaimsKG: Dataset containg factchecked claims from the internet, including tools for finding claims for specific topics 

  • GESIS-Training offers workshops and seminars for collecting and analyzing digital behavioral data  

  • Consulting for research projects using social media data, sensor data or webtracking.