Home: the most loved dishes at each dining hall, with search and diet filters
Near me: what's being served at the closest halls, matched to your restrictions
Dish page: description, allergens, ingredients and nutrition facts
Share of items that fit each diet, by dining hall
Stats: how the dining halls compare on dietary coverage this week
Anonymous feedback: rate a meal or request a new dish
Dining hall forums: anonymous threads for each hall
GIF
Posting anonymously in a dining hall forum
GIF
Filtering dishes by dining hall, meal, diet and allergen
GIF
Leaving a comment on a dish
The BU Big Backs logo. Big eats, big backs!

Inspiration

BU's dining hall menus live on separate sites in different formats. That makes it hard to find a meal that fits your diet, especially if you have allergies or dietary restrictions. We wanted one place where students could see everything being served, filter it to what they can actually eat, and find out what other students love. Big eats, big backs!

What it does

BU Big Backs pulls the menus from the three Fresh Food Co. halls (West Campus, Marciano Commons and Warren Towers) into one app.

Search and filter every dish on campus by name, hall, meal, station, diet (vegetarian, vegan, gluten-free, halal, Cool Food) or allergen
Dish details: description, ingredients, nutrition facts, allergens and student comments
Most loved dishes: tap the heart to like a dish, and the home page ranks the favorites
Near me: what's being served at the closest halls, matched to your restrictions
Dining hall stats: compares halls on dietary coverage and nutrition over time, so dining staff can spot gaps
Anonymous feedback and forums: rate a meal, comment on a dish, request a new one, or post in each hall's thread
Profile: save your dietary restrictions, allergies, home hall and meal plan, with optional Google sign-in

How we built it

Frontend: TanStack Start, React 19, TypeScript, Tailwind CSS and shadcn/ui, with file-based routing (one file per page)
Database and auth: Supabase (Postgres + Google OAuth), with migrations managed in Drizzle
Backend: a Python standard-library HTTP server that serves a JSON API (/api/menu, /api/items, /api/stats, /api/metadata)
Data: BU's public Dining API. A poller pulls a 7-day menu window on startup and once a day after that. We turn the menu snapshots into pandas dataframes and use pandas and numpy to build the per-hall diet and nutrition summary behind the stats page.
Dish photos: generated with a script so every dish has an image

Challenges we ran into

Each hall's menu data comes in a different shape, so we had to normalize it into one format before we could search or filter across halls.
Keeping menus fresh without hammering the API meant building a daily poller and caching a week of menus locally.
Wiring together a TypeScript frontend, a Python backend and Supabase auth, and getting them to talk cleanly in development.

What we learned

How to clean and analyze real-world data with pandas and numpy
How to design a small JSON API and connect it to a modern React frontend
How to set up OAuth and a hosted Postgres database with Supabase
That good data can help both sides: students find meals that fit their needs, and dining staff see where halls fall short