FoodieTrack

Home Page
User Interaction Page
Favourite Cuisine Selection
AI Analysis

Inspiration

Starting university is overwhelming with all the new classes, new routines, and for many first‑years, the first time making daily food decisions without any guidance. We noticed first-hand how often students struggled to choose what to eat while juggling dietary restrictions, fitness goals, emotional well‑being, and the stress of residence life. We wanted to build something that doesn’t just list menu items, but actually understands mood, history, and dietary needs to make food choices easier and healthier.

What it does

Recommends what to eat from dining halls around campus for first-year Waterloo students living in residence, based on audio recordings, pre-selected preferences, diet restrictions, diet goals, physical and emotional feelings, and past data.

How we built it

React + JavaScript on the frontend to collect user input through clean, responsive UI components. Python on the backend to orchestrate logic, data processing, and API communication. Multiple AI and LLM services to interpret audio, extract mood and context, and generate meaningful recommendations. A structured pipeline that converts all user audio, preferences, and history into clean data that Gemini can process effectively.

Challenges we ran into

Integrating the frontend and backend into a smooth, unifying them. Converting messy, multi‑modal inputs (audio, text, preferences) into structured data that Gemini can reliably interpret. Managing multiple APIs and ensuring they worked together under hackathon time pressure.

Accomplishments that we're proud of

Successfully implementing our full project idea from concept to working prototype. Building a functional pipeline that uses real audio input and AI‑powered analysis. Creating a polished, usable interface that first‑year students could realistically adopt. Developing something that genuinely helps new students adjust to campus life and make healthier, more confident food choices.

What we learned

How to use Snowflake as a structured dataset and storage solution. Effective Gemini prompting for emotional and contextual understanding. Implementing Auth0 for secure, reliable user authentication. How to design a multi‑modal AI system that balances usability, accuracy, and speed.

What's next for FoodieTrack

Expanding recommendations to all restaurants within a certain vicinity using Google Maps integration. Publishing FoodieTrack so first‑year students can use it during orientation and throughout the year. Adding real‑time menu updates, nutrition breakdowns, and habit‑tracking features.