Privacy-Compliant User Characterization in Enterprise Social Networks
The thesis designs, implements, and evaluates a framework that turns raw interaction metadata into fuzzy behavioral representations and groups users into overlapping, interpretable segments. Privacy is handled by design through consent handling, pseudonymization, and generalized attributes.
The work connects the Beekeeper industry case studies (ICEDEG 2024, Knowledge 2022) with a reusable open-source Python package and an SSRN framework paper.