Created 23 May 2026 by vesahaaparanta
LunchLens is a campus lunch forecasting tool that combines tabular ML and LLM menu scoring to predict daily visitor counts from history, weather, and competitor menus. It compares a fixed baseline prep policy (70% of max capacity) against a data-driven model, tracking wasted food in kilograms and total cost in euros—including shortage penalties and cumulative savings over time. The repo ships a runnable FastAPI + React MVP with synthetic data, an interactive dashboard, and an open benchmark for food-waste reduction research.
