ImpactQuest — Gamify Social Good for Gen Z

Inspiration

"Gen Z wants to change the world, but traditional volunteering feels outdated."

We identified a critical paradox: Gen Z is statistically the most socially conscious generation in history, yet volunteer participation rates continue to decline. Our research revealed why the current model of community service fails to resonate:

Friction-heavy onboarding — Paper forms and institutional sign-ups kill momentum
Disconnected impact — Volunteers rarely see the tangible results of their contribution
Delayed gratification — In an era of instant feedback, waiting weeks for recognition feels archaic
Scale blindness — "What difference does my one hour make?" remains unanswered

The insight: What if community service delivered the same psychological rewards as gaming? What if helping your neighborhood triggered the same dopamine response as leveling up in your favorite RPG?

That's when ImpactQuest was born — a platform that transforms real-world social impact into an engaging, location-based experience.

What it does

ImpactQuest is a Progressive Web App (PWA) that gamifies community service through location-based quests, AI-powered verification, and social competition.

Core Features

Feature	Description
Quest Discovery	Interactive map displays nearby micro-volunteering opportunities — tree planting, litter cleanup, elderly assistance
AI Photo Verification	Computer vision validates quest completion through proof photos in under 2 seconds
XP & Leveling System	Earn experience points, unlock achievement badges, and climb community leaderboards
Streak Rewards	Consecutive daily completions multiply rewards, building sustainable volunteering habits
Community Leaderboards	Compete with friends, educational institutions, or city-wide rankings
Impact Dashboard	Real-time analytics visualizing collective community contribution

The Quest Loop

Discover Quest → Accept Challenge → Complete Task → 
Submit Photo Proof → AI Verification → Earn XP → Level Up → Repeat

Quest Categories

Environment — Tree planting, neighborhood cleanups, recycling initiatives
Community — Neighbor assistance, food drives, local event support
Education — Tutoring sessions, library volunteering, youth mentoring
Health — Blood donation drives, wellness programs, fitness events
Animal Welfare — Shelter volunteering, stray animal care, adoption drives

How we built it

Architecture Overview

┌─────────────────────────────────────────────────────────┐
│                    FRONTEND (PWA)                        │
│  Next.js 16 + TypeScript + Tailwind CSS + Leaflet Maps  │
└─────────────────────┬───────────────────────────────────┘
                      │
                      ▼
┌─────────────────────────────────────────────────────────┐
│                   BACKEND (Serverless)                   │
│         Next.js API Routes + Edge Functions              │
└─────────────────────┬───────────────────────────────────┘
                      │
          ┌───────────┴───────────┐
          ▼                       ▼
┌─────────────────┐    ┌─────────────────────┐
│    SUPABASE     │    │   HUGGING FACE AI   │
│  PostgreSQL DB  │    │  Image Verification │
│  Auth + Storage │    │   BLIP/ViT Models   │
└─────────────────┘    └─────────────────────┘

Technology Stack

Layer	Technology
Frontend	Next.js 16 (App Router), TypeScript, Tailwind CSS
Maps	Leaflet.js with OpenStreetMap
Backend	Next.js API Routes (Serverless Edge Functions)
Database	Supabase PostgreSQL with Row Level Security
Authentication	Supabase Auth (OAuth + Magic Link)
File Storage	Supabase Storage for proof photos
AI Verification	Hugging Face Inference API (BLIP + ViT models)
PWA	Web App Manifest + Service Workers
Deployment	Vercel Edge Network

AI Verification Pipeline

# Simplified verification logic
def verify_quest_completion(image, quest_criteria):
    # 1. Image captioning with BLIP
    caption = blip_model.generate_caption(image)

    # 2. Object detection with ViT
    detected_objects = vit_model.detect(image)

    # 3. Match against quest requirements
    confidence = calculate_match(caption, detected_objects, quest_criteria)

    return confidence > 0.7  # 70% threshold

Our AI verification system analyzes:

Scene context — outdoor environments, parks, streets, community spaces
Objects present — trash bags, plants, cleaning equipment, people
Actions implied — cleaning, planting, helping, organizing

Challenges we ran into

Challenge 1: AI Verification Accuracy

Problem: Initial AI models struggled to distinguish between "holding a trash bag" versus "actively cleaning up an area."

Solution: We implemented a multi-model ensemble approach:

BLIP model for holistic scene understanding and captioning
ViT model for granular object detection
Custom scoring algorithm weighing multiple visual factors
Fallback mechanism to manual review for edge cases

Challenge 2: Location Spoofing Prevention

Problem: Users could potentially fake GPS coordinates to complete quests remotely without physical presence.

Solution: We implemented multiple verification layers:

Cross-referencing photo EXIF metadata with claimed quest location
Requiring photos to be captured within the app (blocking gallery uploads)
Velocity checks preventing physically impossible quest completions

Challenge 3: Real-time Leaderboard Performance

Problem: Calculating rankings across thousands of concurrent users caused significant database bottlenecks.

Solution: We optimized the data layer:

PostgreSQL materialized views for leaderboard caching
Optimized RPC functions with proper indexing strategies
Incremental XP updates instead of full recalculation on each action

Challenge 4: PWA Compatibility on iOS Safari

Problem: iOS Safari has limited PWA support, causing geolocation and notification issues.

Solution: We implemented graceful degradation:

Browser-specific polyfills for unsupported features
Alternative manual location input for problematic browsers
Clear, user-friendly permission request flows

Accomplishments that we're proud of

Technical Achievements

Sub-2-second AI verification — Users receive instant feedback on quest completion
100% serverless architecture — Infinite scalability with zero idle costs
Offline-capable PWA — Background sync ensures no data loss on poor connections
Real-time updates — Leaderboard changes reflect instantly via Supabase subscriptions

Design Achievements

Premium minimalist UI — Clean white-violet color scheme optimized for Gen Z aesthetics
Micro-interactions — Satisfying animations reinforce positive behavior
Mobile-first responsive design — Native app feel in a progressive web app
Accessibility compliance — WCAG 2.1 AA compliant color contrasts and navigation

Projected Impact

We modeled the potential community impact assuming ImpactQuest reaches 10,000 active users:

$$\text{Monthly Impact} = \underbrace{10,000}{\text{users}} \times \underbrace{4}{\text{quests/week}} \times \underbrace{4}_{\text{weeks}} = 160,000 \text{ micro-actions}$$

That translates to 160,000 trees planted, trash bags collected, or neighbors assisted — every single month from a single city deployment.

What we learned

Technical Learnings

Edge functions eliminate cold starts — Serverless no longer means slow initial responses
Supabase RLS simplifies security — Row-level security policies replace complex middleware auth
AI verification requires ensemble approaches — Single models fail on edge cases; combining multiple models dramatically improves accuracy
PWAs can match native experiences — With proper optimization, web apps feel indistinguishable from native

Product Learnings

Gamification requires careful balance — Rewards too easy feel meaningless; too difficult causes abandonment
Social proof drives engagement — Visible leaderboards increase completion rates by approximately 3x
Instant feedback is non-negotiable — Modern users abandon flows that require manual approval delays
Location features require trust — Privacy-first design is essential for location-based applications

Personal Growth

Building under extreme time pressure forces ruthless feature prioritization
"Perfect is the enemy of shipped" — launching imperfect but functional beats polished but incomplete
The best features are the ones users actually use, not the most technically impressive

What's next for ImpactQuest

Immediate Roadmap (Q1 2026)

Feature	Status
Team Quests — Form squads and compete collaboratively	In Development
Corporate Partnerships — Sponsored quests with tangible rewards	Planned
Educational Integration — Track community service hours for academic transcripts	Planned
Verified Impact Certificates — Blockchain-backed proof of contribution	Research Phase

Long-term Vision

"The Strava of Social Good"

Just as Strava transformed running into a social, competitive activity, ImpactQuest aims to make community service the next mainstream social movement.

Our vision includes:

City-wide leaderboards visible to all residents
Educational institutions competing in annual "Impact Week" challenges
Corporate CSR programs integrated with employee engagement tracking
A global real-time map visualizing acts of community service worldwide

Sustainability Model

┌─────────────────────────────────────────┐
│  B2C: Free forever for individuals      │
├─────────────────────────────────────────┤
│  B2B: Enterprise dashboard for:         │
│    • Corporations (CSR tracking)        │
│    • Schools (service hour credits)     │
│    • NGOs (volunteer management)        │
├─────────────────────────────────────────┤
│  Sponsored Quests: Brands sponsor       │
│  featured challenges with real prizes   │
└─────────────────────────────────────────┘

Conclusion

ImpactQuest represents a fundamental reimagining of how we approach community service.

Rather than guilt-driven appeals, we meet users where they already are: on their phones, seeking meaningful engagement, wanting to contribute to something larger than themselves.

Every quest completed. Every experience point earned. Every level achieved.

It all translates to measurable, real-world community impact.

Built with dedication for the Supabase Open Source Hackathon

"Making the world better, one quest at a time."