SynEdu-HEDL: A Synthetic Dataset for Early Warning Prediction of Student Success

SynEdu‑HEDL: A Synthetic Dataset for Early Warning Prediction of Student Success This repository hosts SynEdu‑HEDL, a fully synthetic dataset that mimics real‑world student behaviors, academic performance, and digital engagement in higher education. Designed to support privacy‑preserving research in learning analytics and educational data mining, it includes six interconnected tables (student profiles, course metadata, LMS logs, assessments, engagement metrics, and outcome labels) with realistic correlations. The dataset enables development and benchmarking of machine learning models for early prediction of at‑risk students, dropout risk, and learning gains – all without exposing sensitive real student data.