WashU Meal Plan Optimizer (based on 2026-27 meal point policy)

Predict your next-semester WashU Dining spending from local uploads, then choose the cheapest practical meal plan instead of overbuying points.

Assumption: if a student starts living off campus, the model assumes they will not move back to a dorm later in the academic year.

1) Choose Prediction Mode

If you are going to deside your plan for fall semester, choose fall mode; if you are going to deside your plan for spring semester, choose spring mode.

2) Spring Student Context

Spring mode keeps its own student context and optional rollover inputs.

Required spring inputs

Meal transactions CSV only. On-campus students use the full-day campus assumption.

3) Spring Inputs

Manual on-campus time replaces schedule file uploads.

Living on campus is set to Yes, so schedule input is skipped and the model assumes full-day campus presence.

Fix the spring input issues below to enable analysis.

  • - Meal transactions CSV: no file selected

4) Spring Results

Prediction output and plan recommendation dashboard for spring mode.

Add your meal CSV and required on-campus time inputs, then run analysis to view results.

5) Methodology & Caveats

  • - Meal transactions are used to measure your real spending pattern: how often you buy lunch, snacks, and dinner, how much you usually spend, and how that changes week to week.
  • - On-campus time inputs are used to estimate when you are likely to need food on campus. Lunch-time presence, evening presence, and long daytime gaps all affect predicted future demand.
  • - In spring mode, the model uses your previous fall behavior to estimate next spring. In fall mode, it uses your previous fall and current spring behavior together to estimate next fall.
  • - If you lived on campus during a semester, the model assumes full-day campus presence for that semester, so schedule input is not required.
  • - The recommendation compares eligible meal plans against predicted spending, rollover rules, likely leftover waste, and reasonable out-of-pocket shortfall to find the cheapest practical option.
  • - Results are directional estimates, not guarantees. Unusual habits, travel, guests, or major routine changes can make actual spending differ.

6) Privacy

All processing runs locally in your browser. No login, no backend database, and no server-side storage. Uploaded files are not persisted; refreshing the page clears all loaded data.

7) Support The Project

If you found this useful, please give the project a star on GitHub.