Hoda Soltani (OUHSC)

Transfer students represent a significant and academically diverse population whose successful integration into four-year institutions requires timely, evidence-based support. Framed within a Data Science for Social Good approach, this session presents a predictive analytics framework developed at the University of Oklahoma to examine transfer students’ prior academic pathways and first-term university outcomes across multiple colleges.

Using principles from educational data mining, I analyze a multidimensional dataset that includes sociodemographic characteristics, prior academic performance, enrollment intensity, financial aid, campus employment, and early indicators of academic engagement. The modeling pipeline applies supervised learning for retention classification alongside explainable machine learning techniques to reveal the drivers of student success and risk.

Key methodological challenges—such as class imbalance, overfitting, and domain-informed feature engineering—are addressed through practical implementation strategies. Most importantly, the session demonstrates how predictive insights can be translated into targeted, personalized interventions that advance student success and institutional effectiveness.