SocialRec: Personalizing Educational Experience through Social Network-based Recommender Systems

This project was undertaken as part of Universidad del Desarrollo's personalization strategic project. It integrated data from multiple sources and formats to build early risk profiles and recommender systems for improving the educational experience.

Privacy and security were central because the data were sensitive. To address this, we developed anonymization and de-identification protocols that preserve enough granularity for analysis while protecting individuals. The team also developed an algorithmic benchmark for inferring social networks from router connections and combined individual, mixed, and relational data to create recommendations based on machine learning and interaction networks.

A project essay is available on Medium.