Please reach out to us if you are interested in learning more about any of these projects. Results will not be made public due to FERPA and other privacy concerns.
Prediction of At-Risk Upperclassmen Students
This project assigns a risk score to students entering their senior year who are likely to discontinue, with predictions provided for each department.
– Independent Variables: Demographics, term GPA, course grades, financial aid status (Pell, SEOG, federal loans), total credit hours earned, and student attribute groups (e.g., athletes).
– Dependent Variables: Enrollment Status of the first semester of their senior year at the 5-week mark.
– Models Used: Random Forest, K-Nearest Neighbors (KNN), Logistic Regression, XGBoost, LightGBM, and Support Vector Machine (SVM). Feature importance was also obtained to identify the most influential variables in the predictions.
Example of feature importance for civil engineering students in predicting senior year enrollment status
Toxic/Synergic Course Combination Impact Analysis
This project analyzes how taking certain courses together affects student pass rates compared to taking each course individually. Course combinations with higher pass rates are identified as synergistic combinations, while those with lower pass rates are labeled as toxic combinations. Although the analysis is preliminary, it has the potential to serve as a valuable tool for advising at-risk students.
Example of Calc 3 pass rate alone and when taken with other courses in the same semester
