CourseCast helps you get better results from Wharton's CourseMatch by showing you how your utility inputs translate into actual schedules, taking into account historical price patterns and variability.
CourseCast runs Monte Carlo simulations in your browser to predict possible schedule outcomes from the CourseMatch system. Each simulation uses a model that predicts closing prices for courses, then solves a linear programming problem to determine which classes get selected.
After running multiple simulations, CourseCast aggregates the results to show you:
- Probability of obtaining each course based on your utility allocations
- Likely schedule configurations you might end up with
- Trade-offs between different utility distributions
This allows you to iteratively refine your course utilities and see how changes affect your chances before submitting to the real CourseMatch system.
The main application is a Next.js web app located in the web/ directory. See web/README.md for setup instructions.
web/- Next.js web application (main app)data/- Historical data exports (preserved for reference)