match-a

Strengthening BC's nonprofit workforce through smart matching, crisis mobilization, retention prediction, and knowledge continuity.

Built for youCode 2026 — SAP Challenge

🔗 Live Demo

The Problem

BC has 29,000 nonprofits contributing $6.7B to the economy. Volunteerism hasn't recovered to pre-pandemic levels and the gap is widening. Organizations are cutting programs because they don't have the people, and when volunteers do leave, their institutional knowledge disappears with them.

Our Solution

match-a is a four-layer volunteer lifecycle platform:

Match — Volunteers complete an onboarding quiz (skills, languages, interests, availability). A weighted compatibility algorithm scores every opportunity against their profile.

Mobilize — When a nonprofit hits a critical gap, they post an urgent request. Matched volunteers receive real-time alerts via Supabase Realtime. One tap to accept.

Retain — A churn prediction engine extracts engagement features (days since active, hours trend, shift frequency, crisis response rate) and classifies volunteers as active, cooling, or at-risk. Coordinators see suggested check-in actions, not raw data.

Preserve — When volunteers leave, a handoff document captures their key contacts, recurring tasks, and tips. The next volunteer inherits it.

Tech Stack

• Frontend: Next.js, TypeScript, Tailwind CSS
• Backend/Auth/DB: Supabase (PostgreSQL, Auth, Realtime)
• Matching: Weighted multi-variable compatibility scoring
• Retention: Rule-based churn classifier with feature extraction pipeline
• Hosting: Vercel

What We Learned

• First time building with Supabase Realtime subscriptions for live crisis alerts
• Designed and implemented a weighted matching algorithm with 5 scoring dimensions
• Built a churn prediction pipeline: feature extraction → classification → actionable output
• Learned role-based auth flows with Supabase Auth
• One of our team member's first time working with full stack development

Design Constraints Addressed

• No IT support: Coordinator goes from signup to posting a request in under 3 minutes. No configuration.
• Accessibility: Language matching built into the data model. Onboarding uses tappable tags, not forms.
• Sustainability: Rule-based matching with no model decay. Runs identically on day 1 and day 180.
• Trust & safety: Background check field gates sensitive volunteer roles.
• Don't rebuild what exists: Platforms like Better Impact handle scheduling. We fill the gap nobody covers — acute crisis mobilization.

Team

• Nicholas Huang
• Shlok Lande
• Emily Tu