Echoes

Welcome Page
Sample Messages
Desk Information and Message Box
About the product page

Inspiration Our inspiration stems from improving university study life by creating a safe community to share course advice, interact with others, and motivates students to be more productive. As students, we also often experience burn out, forget to take breaks and look after ourselves. By implementing Echoes, it provides a short, convenient activity built directly into university desks, encouraging students to reconnect with both themselves, and the student community around them.

What it does: Scan a QR on any desk, read anonymous messages from students who studied there before you, leave one for whoever sits there next. Messages are tagged and surfaced based on your courses and study context so what you read feels written for your exact situation, not just anyone.

How we built it Next.js and Supabase for the core stack, pgvector for embedding-based message ranking, OpenAI's Moderation API for moderation and tagging, and Vercel for deployment. QR codes tied to location IDs in the database. The ranking system uses semantic similarity, course overlap, temporal relevance, and upvotes to surface the most relevant messages for each student's context.

Challenges we faced Honestly, getting the ranking to feel personal was rather challenging. We had to learn about how vectors were used to match relevant study tip messages to the user. Furthermore, drafting the content of the pitch proved to be challenging as we needed plenty of discussion about what the judges would want to see and what extra miles we could take.

Accomplishments that we're proud of The design — it doesn't look or feel like a hackathon project. The moment when switching between two student profiles visibly reorders the same desk's messages. Furthermore, we are proud of implementing the semantic matching ranking to surface relevant study tip messages to the user.

What we learned How to make an appealing web UI design. Incorporating the right balance of colours and pictures to make the website visually interesting and not overwhelming will prove useful in attracting users when adding features for this project and future projects.

What's next for Echoes After refining the code, we plan to run a pilot test in UNSW’s Main Library and gauge user feedback. Furthermore, we plan to add moderation support for more languages to make our website safer for students from all backgrounds.