The Amala Map

Built for the Lovers of Amala by the lovers of Amala

Comment

Built for the lovers of Amala by the lovers of Amala

Inspiration

The Amala Discovery Platform was born from a simple observation: finding authentic, quality Amala spots in Lagos shouldn't be a guessing game. As someone passionate about Nigerian cuisine, I noticed that food enthusiasts often rely on word-of-mouth or spend hours searching through fragmented online reviews to discover great local eateries.

The inspiration struck during a conversation about how technology could preserve and promote traditional Nigerian food culture. I envisioned a platform that would not only help people find amazing Amala restaurants but also celebrate the rich culinary heritage of Lagos through an intelligent, community-driven approach.

What it does

The Amala Map is an intelligent location discovery platform that revolutionizes how people find authentic Amala restaurants across Lagos, Nigeria. The platform combines cutting-edge AI technology with community-driven content to create the most comprehensive Amala restaurant directory.

Core Features:

  • Interactive Google Maps Interface - Browse 38+ verified Amala locations with custom markers showing real-time open/closed status, price ranges, and ratings
  • Autonomous Discovery System - AI-powered background service that automatically discovers new Amala spots from Google Places API, web sources, and social media mentions
  • AI-Powered Submissions - Natural language chat interface using Google Gemini AI that extracts structured location data from conversational descriptions like "There's this amazing Amala spot near Ikeja with great pepper sauce"
  • Real-time Moderation Panel - Admin dashboard with live notifications for reviewing and approving community submissions with one-click actions
  • Advanced Filtering - Smart filters for open hours, price range, service type (dine-in/takeaway), and cuisine specialties
  • Mobile-First Design - Responsive interface optimized for mobile discovery with touch-friendly interactions and bottom sheet navigation
  • The Magic: Users can simply chat with the AI about Amala spots they know, and the system intelligently extracts location details, validates them against existing data, and queues them for community review - making it effortless to contribute to Lagos's food culture preservation.

How we built it

Architecture & Tech Stack: We built The Amala Map using a modern, scalable architecture centered around Next.js 15 with TypeScript for type safety and developer experience.

Frontend Development:

  • Next.js 15 + App Router for server-side rendering and optimal performance
  • Tailwind CSS + Radix UI for accessible, responsive design components
  • Google Maps JavaScript API with custom marker clustering and info windows
  • Real-time state management using React hooks with optimistic updates

AI Integration:

  • Google Gemini Pro for natural language processing in location submissions
  • Custom AI service layer that extracts structured data from conversational input
  • Confidence scoring algorithm that validates AI-extracted information against known patterns
  • Fallback extraction system when AI services are unavailable

Autonomous Discovery Engine:

// Multi-phase discovery process
const discoveryPhases = [
  'Google Places API Search',    // Primary source
  'Web Scraping Analysis',       // Secondary validation  
  'Social Media Monitoring',     // Community insights
  'AI Validation & Scoring'      // Quality assurance
];

Database & Backend:

  • Supabase PostgreSQL with real-time subscriptions for live moderation updates
  • Custom database schema optimized for geospatial queries and duplicate detection
  • API routes handling autonomous discovery, moderation workflows, and location management
  • Environment-based configuration with graceful degradation when services are unavailable

Development Process:

  • Core map interface and location display system
    • AI integration and natural language submission flow
    • Autonomous discovery system and moderation panel
    • Production optimization, mobile responsiveness, and deployment

Challenges we ran into

  1. AI Hallucination & Data Accuracy Challenge: Google Gemini occasionally generated plausible but incorrect location details, creating fake restaurants that seemed real.

Solution: Implemented a multi-layer validation system:

  • Cross-validation with Google Places API for address verification
  • Confidence thresholds requiring 80%+ certainty before auto-approval
  • Human moderation queue for all AI-generated submissions
  • Similarity detection to prevent duplicate entries
  1. Real-time Updates Without WebSocket Complexity Challenge: Needed live moderation notifications and map updates but wanted to avoid WebSocket infrastructure complexity.

Solution: Developed an optimistic update pattern:

// Immediate UI feedback
setLocations(prev => [...prev, newLocation]);
// Background sync with rollback on failure
try {
  await saveLocation(newLocation);
} catch (error) {
  setLocations(prev => prev.filter(loc => loc.id !== newLocation.id));
}
  1. Mobile Performance with Large Datasets Challenge: Rendering 50+ map markers caused significant lag on mobile devices, especially older phones.

Solution: Implemented viewport-based clustering and lazy loading:

  • Only render markers visible in current map bounds
  • Cluster nearby locations when zoomed out
  • Progressive loading as users pan and zoom
  • Optimized marker icons using SVG data URLs
  1. Autonomous Discovery Reliability Challenge: Initial web scraping returned many false positives and irrelevant locations.

Solution: Built a sophisticated filtering pipeline:

  • Keyword matching for Amala-specific terms in multiple languages
  • Geospatial validation ensuring locations are within Lagos boundaries
  • Business hours and contact information verification
  • Community validation through the moderation system
  1. Graceful Degradation Strategy Challenge: External API failures could break the entire user experience.

Solution: Comprehensive fallback systems:

Accomplishments that we're proud of

🤖 Autonomous Discovery Success

  • Built fully functional AI system that discovered 38+ real Amala locations across Lagos
  • Achieved 85% accuracy in location validation with minimal false positives
  • Created a self-improving system that gets better with community feedback

🚀 Production-Ready Performance

Sub-2 second page loads on mobile devices 100% responsive design working flawlessly across all screen sizes

Built With

Share this project:

Updates