May 2025
Given a variety of academic and environmental factors, my small team and I predicted students' exam performance through feature engineering, hierarchical cross-validation, principal component analysis, random forest, and linear regression. We visualized our insights using principal component analysis biplots with loadings, correlation heatmaps, box plots, and distribution charts.