This R project is designed for education analytics at the Vrije Universiteit Amsterdam (VU). It focuses on predicting the conversion of applicants to enrolled students using a Random Forest model.
- Reads in and preprocesses data on student applications and enrollments
- Calculates historical conversion rates for different student groups
- Trains a Random Forest model to predict student enrollment based on application data
- Generates predictions for the current academic year
TBD
- Ensure you have the necessary system variables set in your .Renviron file.
- Run the R script to generate the enrollment predictions.
The project uses the following key libraries and techniques:
vusa: A vusaverse packagerenv: For managing package dependencies- Random Forest modeling with the
rangerpackage - Feature engineering, including handling of missing data and creating lagged conversion rates