Synthetic Educational Datasets for Fair and Transparent AI in Learning Analytics

This repository contains two synthetic educational datasets generated for the research paper “Towards Fair and Transparent AI: Leveraging Synthetic Data for Equity in Learning Analytics.”The datasets replicate realistic student academic performance and interaction data to support fairness-aware machine-learning experiments in educational settings.

All data were created synthetically using Python libraries such as SDV, CTGAN, and SynthPop, ensuring that no real student information is included.The datasets are designed to facilitate reproducible research on bias mitigation, model interpretability, and equitable learning analytics without compromising data privacy.

Contents:

synthetic_dataset1_academic.csv — Simulated records of student performance across subjects.

synthetic_dataset2_interaction.csv — Simulated learning-management interaction data (e.g., quiz attempts, submissions, and activity frequency).

README.md — Detailed description of variables, generation process, and usage instructions.

Usage:These datasets may be freely used for research and educational purposes, provided proper citation to the above paper and dataset DOI is given.

Citation:Joshi, Yogesh. (2025). Synthetic Educational Datasets for Fair and Transparent AI in Learning Analytics [Data set]. Zenodo. DOI: (to be added after upload)