Inspiration
Restaurants often overprep to avoid stockouts, but that leads to wasted food, labor, and cost. We wanted to build a tool that helps kitchens prep closer to real demand.
What it does
WasteLess forecasts demand using restaurant sales data, weather, and local events. The goal is to reduce overproduction and food waste.
How we built it
We built WasteLess with React + TypeScript on the frontend and Python + FastAPI on the backend. We used pandas for data processing, joblib for model loading, and external event data sources like the Ticketmaster API to add demand context.
Challenges we ran into
Our biggest challenge was cleaning and organizing multiple datasets. Another was incorporating external signals like weather and events in a way that actually improves forecasting.
Accomplishments that we're proud of
We built a full-stack prototype tied to a real operational problem. It includes a polished frontend, CSV upload flow, and a forecasting pipeline that goes beyond just historical sales.
What we learned
We learned that context matters as much as the model. Clean data pipelines, useful signals, and clear product design all make a big difference.
What's next for WasteLess
Next, we want to improve the model, strengthen the live events pipeline, and generate more actionable prep recommendations for real restaurant use.
Built With
- fastapi
- pandas
- python
- react
- typescript
Log in or sign up for Devpost to join the conversation.