SocialRec and educational recommender systems

Personalization in education is often treated as a content problem: recommend the next resource, video, or exercise. SocialRec approached personalization as a relational and institutional problem.

The project integrated heterogeneous university data, privacy-preserving protocols, and network inference to understand educational experience beyond individual records. A central question was how interaction patterns, campus routines, and student trajectories can inform early support without exposing sensitive personal information.

The technical work combined anonymization, de-identification, social network inference from router connections, mixed individual and relational data, and machine learning. The applied goal was a recommender system that helps universities identify useful support actions, not just predict risk.

SocialRec connects directly with CRiSS-LAB research on learning analytics, higher education systems, collective intelligence, and network-based recommendation.

Read the project essay on Medium.