Interdisciplinary Approaches to User Modeling in Data-Driven Systems

This PhD project proposes a rethinking of user modeling in data science by developing dynamic, context-aware, and interdisciplinary approaches that better reflect the complexity of human behavior in digital environments. Current computational models often rely on inherited assumptions from early psychological, communicative, or economic theories; assumptions that portray users as static, isolated, and predictable. Although such frameworks are outdated, they remain embedded within algorithms and system architectures, where they are rarely made explicit and are silently propagated at scale.

By drawing on insights from the social sciences and humanities, this research enriches core data science methods, such as machine learning, natural language processing, and network analysis, with more nuanced understandings of human interaction, communication practices, and social context. The project aims to expose and critique the hidden theoretical legacies within existing models while constructing alternatives that are flexible, adaptive, and better aligned with real-world behavior.

Through applied studies in areas such as social platforms, recommender systems, and interactive technologies, the project shows how interdisciplinary, value-aware modeling can lead to both improved system performance and more responsible representations of users. Ultimately, it contributes to the development of a new generation of user models that advance current practice in data science while supporting more transparent, accountable, and human-centered system design, aligned with the principles of Digital Humanism.

This project is supervised by faculty at TU Wien, WU Wien, and University of Vienna.