LafaekStreet Public Dashboard

Live: https://www.lafaekstreet.com

A cutting-edge civic engagement platform that empowers citizens of Timor-Leste to report infrastructure issues and track their resolution through an intelligent public dashboard powered by Amazon Nova 2 Lite AI.

🎯 AI-Powered Civic Engagement

This project showcases the power of Amazon Nova 2 Lite in transforming civic engagement through:

Intelligent Chat Assistant (RuaLafaek) - Natural language queries about infrastructure data
Real-time Data Analysis - AI-powered insights from citizen reports
Smart Query Processing - Converting natural language to database queries
Automated Response Generation - Context-aware responses with suggestions

🚀 What's Already Implemented

✅ Core Platform Features

🏠 Dashboard Homepage

Dashboard Homepage

Real-time Statistics Cards: Total reports, pending, in-progress, and fixed counts
Interactive Map Preview: Clustered markers showing report locations
Performance Metrics: Response times, resolution rates, and system health
Recent Activity Feed: Live updates of platform activity
Quick Action Navigation: Direct access to all major features

🗺️ Interactive Map System

Interactive Map

Full-screen Map Interface: Leaflet-based mapping with OpenStreetMap tiles
Smart Marker Clustering: Automatic grouping of nearby reports for better performance
Advanced Filtering: Filter by status, issue type, municipality, and date ranges
Color-coded Status Indicators: Visual distinction between pending, in-progress, fixed, and rejected reports
Detailed Popups: Rich information display for each report marker
Responsive Design: Optimized for both desktop and mobile viewing

📊 Comprehensive Statistics

Statistics Page

Multi-dimensional Analytics: Reports by type, municipality, and time series
Interactive Charts: Pie charts, bar charts, and line graphs using Recharts
Performance Tracking: Resolution rates, average response times, and trends
Municipality Breakdown: Detailed statistics for each region in Timor-Leste
Time Series Analysis: 30-day trend visualization with submission patterns

📋 Report Management

Reports List

Advanced Search & Filtering: Multi-criteria filtering with real-time updates
Pagination System: Efficient handling of large datasets
Detailed Report Views: Complete report information with images and status history
Status Tracking: Visual indicators for report lifecycle stages
AI-Powered Similarity Search: Vector-based similar report detection using Amazon Nova 2 Multimodal Embeddings
Mobile-optimized Interface: Responsive design for all screen sizes

Report Detail

🤖 RuaLafaek AI Chat Assistant (Amazon Nova 2 Lite Integration)

RuaLafaek AI Chat

Core AI Capabilities

Natural Language Processing: Understands citizen queries in conversational language
Intelligent Query Classification: Automatically categorizes questions as database queries, general info, or help requests
Dynamic SQL Generation: Converts natural language to safe PostgreSQL queries
Context-aware Responses: Provides relevant, formatted answers with follow-up suggestions
Multi-step AI Pipeline: Sophisticated processing chain for optimal results

Security & Protection

hCaptcha Integration: Bot protection with verification system
Input Validation: 500-character limit with safety checks
SQL Injection Prevention: Whitelist-only query validation
Rate Limiting: Protection against abuse and automated requests

User Experience

Floating Chat Widget: Non-intrusive interface with custom AI mascot
Rich Markdown Responses: Formatted text with headers, lists, and emphasis
Suggestion System: Contextual follow-up questions for continued engagement
Error Handling: Graceful fallbacks with helpful error messages
Theme Integration: Seamless light/dark mode support

🔍 AI-Powered Similarity Search

Vector Embedding Technology

Amazon Nova 2 Multimodal Embeddings: Generates 1024-dimensional vectors from report content
Multimodal Analysis: Combines text descriptions and image content for comprehensive similarity
pgvector Integration: PostgreSQL extension for efficient vector operations
Cosine Similarity: Mathematical similarity scoring between 0-1

Smart Detection

Automatic Processing: All reports generate embeddings upon submission
Configurable Thresholds: Default 50% similarity threshold (adjustable 0-100%)
Real-time Search: Find similar reports instantly from any report detail page

User Interface Features

"Find Similar" Button: One-click similarity search from report details
Similarity Scores: Visual percentage indicators showing relationship strength
Smart Filtering: Results filtered by relevance and public visibility
Loading States: Real-time feedback during embedding analysis

Technical Implementation

sql

-- Vector similarity query using cosine distance
SELECT r.*, (1 - (e2.embedding <=> e1.embedding)) as similarity_score
FROM report_embeddings e1
JOIN report_embeddings e2 ON e1.embedding_type = e2.embedding_type
JOIN reports r ON e2.report_id = r.id
WHERE e1.report_id = $1 AND similarity_score >= $2
ORDER BY similarity_score DESC

Use Cases

Citizens: Discover if their issue has been reported before
Government: Identify patterns and recurring problem areas
Analytics: Track similar infrastructure issues across regions
Fraud Prevention: Detect potential duplicate or spam reports

🏗️ System Architecture

🔄 AI Chat Flow Diagram

🛠️ Technology Stack

Frontend

Next.js 14: App Router with server-side rendering and static generation
React 19: Latest React features with concurrent rendering
TypeScript: Full type safety across the application
Tailwind CSS: Utility-first styling with custom design system
Framer Motion: Smooth animations and transitions

AI & Machine Learning

Amazon Nova 2 Lite: Advanced language model via AWS Bedrock
Amazon Nova 2 Multimodal Embeddings: 1024-dimensional vectors for similarity search
LangChain: AI application framework for complex workflows
AWS Bedrock: Managed AI service infrastructure
Natural Language Processing: Query understanding and response generation
Vector Similarity Search: pgvector with cosine similarity for duplicate detection

Backend & Database

PostgreSQL: Robust relational database with advanced features
pgvector Extension: Vector similarity search with 1024-dimensional embeddings
Prisma ORM: Type-safe database access with migrations
Next.js API Routes: Serverless API endpoints
Connection Pooling: Optimized database performance

Maps & Visualization

Leaflet: Open-source mapping library
React Leaflet: React integration for Leaflet
Marker Clustering: Performance optimization for large datasets
Recharts: Responsive chart library for data visualization

Security & Validation

hCaptcha: Bot protection and verification
Input Sanitization: XSS and injection prevention
SQL Safety Validation: Whitelist-based query filtering
Environment Variable Management: Secure configuration handling

🚀 Getting Started

Prerequisites

Node.js 18+
PostgreSQL 14+
AWS Account with Bedrock access
hCaptcha account (optional, for bot protection)

Installation

Clone the repository

bash

git clone https://github.com/ajitonelsonn/lafaekstreet.git
cd lafaekstreet_dashboard

Install dependencies
bash
```
npm install
```
Environment Setup
bash
```
cp .env.example .env.local
```

Configure Environment Variables

env

# Database
DATABASE_URL="postgresql://user:password@localhost:5432/lafaekstreet"

# AWS Configuration (Required for AI Chat)
AWS_REGION="us-east-1"
AWS_ACCESS_KEY_ID="your_access_key"
AWS_SECRET_ACCESS_KEY="your_secret_key"
BEDROCK_MODEL_ID="us.amazon.nova-2-lite-v1:0"

# hCaptcha (Optional)
NEXT_PUBLIC_HCAPTCHA_SITE_KEY="your_site_key"
HCAPTCHA_SECRET_KEY="your_secret_key"

# App Configuration
NEXT_PUBLIC_APP_URL="http://localhost:3000"

Database Setup
bash
```
npm run db:generate
```
Start Development Server
bash
```
npm run dev
```
Open Application Navigate to http://localhost:3000 (local) or https://www.lafaekstreet.com (production)

📁 Project Structure

lafaekstreet_dashboard/
├── app/                          # Next.js App Router
│   ├── api/                     # API Routes
│   │   ├── chat/               # AI Chat endpoint
│   │   ├── statistics/         # Statistics APIs
│   │   ├── reports/            # Report management APIs
│   │   └── health/             # Health check endpoints
│   ├── map/                    # Interactive map page
│   ├── reports/                # Report browsing page
│   ├── statistics/             # Analytics dashboard
│   └── layout.tsx              # Root layout with AI chat
├── components/                  # React Components
│   ├── chat/                   # AI Chat components
│   │   ├── LafaekAIChat.tsx   # Main chat widget
│   │   ├── ChatMessage.tsx    # Message rendering
│   │   └── ChatInput.tsx      # Input with validation
│   ├── map/                    # Map components
│   ├── statistics/             # Chart components
│   ├── reports/                # Report components
│   └── common/                 # Shared components
├── lib/                        # Utilities & Configuration
│   ├── ai/                     # AI Integration
│   │   ├── nova-client.ts     # Amazon Nova client
│   │   ├── database-handler.ts # SQL generation
│   │   └── web-scraper.ts     # General info handler
│   ├── prisma.ts              # Database client
│   └── hcaptcha.ts            # Bot protection
├── prisma/                     # Database Schema
├── public/                     # Static Assets
└── styles/                     # Global Styles

🔌 API Endpoints

Core Data APIs

GET /api/statistics - Dashboard statistics and metrics
GET /api/reports - Paginated reports with advanced filtering
GET /api/reports/[id] - Individual report details
GET /api/reports/[id]/similar - AI-powered similar reports using vector embeddings
GET /api/reports/map - Optimized data for map visualization

AI Chat API

POST /api/chat - Main AI chat endpoint with Nova integration
GET /api/health/db - Database connectivity check

Analytics APIs

GET /api/statistics/by-type - Reports grouped by issue type
GET /api/statistics/by-municipality - Regional breakdown
GET /api/statistics/time-series - Trend analysis data

🎨 Features Showcase

Intelligent Conversational Interface

Natural language queries: "Show me all potholes in Dili"
Context-aware responses with rich formatting
Follow-up suggestions for continued exploration
Multi-language support preparation

Advanced Data Visualization

Real-time dashboard with live statistics
Interactive charts with hover details and animations
Geographic visualization with clustering algorithms
Responsive design optimized for all devices

AI-Powered Similarity Search

Vector-based similar report detection using 1024-dimensional embeddings
Multimodal analysis combining text and image content
Configurable similarity thresholds for precise matching
Real-time duplicate detection and fraud prevention

Comprehensive Filtering System

Multi-criteria search across all report fields
Date range filtering with calendar integration
Geographic filtering by municipality and district
Status and priority-based filtering

Performance Optimizations

Server-side rendering for fast initial loads
Incremental static regeneration for cached content
Database query optimization with indexing
Image optimization and lazy loading

🔒 Security Implementation

AI Chat Security

hCaptcha verification prevents automated abuse
Input length limits (500 characters) prevent overload
SQL injection prevention with whitelist validation
Rate limiting protects against spam attacks

Data Protection

Environment variable encryption for sensitive data
Database connection pooling with secure credentials
HTTPS enforcement in production environments
Input sanitization across all user inputs

🌟 Amazon Nova Integration Highlights

Why Amazon Nova 2 Lite?

Cost-effective: Optimized for high-volume civic engagement use cases
Fast Response Times: 3-8 second average response for complex queries
Reliable Performance: Consistent quality for government applications

AI Pipeline Architecture

Question Classification: Determines query type (database/general/help)
Context Processing: Analyzes user intent and required data
Query Generation: Converts natural language to safe SQL
Response Formatting: Creates markdown-formatted, user-friendly responses
Suggestion Generation: Provides relevant follow-up questions

Real-world Impact

Citizen Empowerment: Easy access to infrastructure data
Government Transparency: Open data with intelligent querying
Scalable Architecture: Ready for nationwide deployment

🚀 Deployment & Production

Hosting: Vercel

Deployed on Vercel at https://www.lafaekstreet.com
Automatic deployments from Git
Environment-specific configuration management
Serverless API routes with edge optimization

Performance Monitoring

Real-time API response tracking
Database query performance analysis
AI model usage and cost monitoring
User engagement analytics

� Project Summary

LafaekStreet demonstrates the transformative power of Amazon Nova 2 Lite in civic engagement, providing citizens of Timor-Leste with an intelligent, accessible platform for infrastructure reporting and government transparency. The AI-powered chat assistant, RuaLafaek, makes complex data queries as simple as asking a question in natural language, bridging the gap between citizens and their government data.

Key Innovation: Converting natural language citizen queries into actionable database insights, making government data truly accessible to everyone.

LafaekStreet Public Dashboard ​

🎯 AI-Powered Civic Engagement ​

🚀 What's Already Implemented ​

✅ Core Platform Features ​

🏠 Dashboard Homepage ​

🗺️ Interactive Map System ​

📊 Comprehensive Statistics ​

📋 Report Management ​

🤖 RuaLafaek AI Chat Assistant (Amazon Nova 2 Lite Integration) ​

Core AI Capabilities ​

Security & Protection ​

User Experience ​

🔍 AI-Powered Similarity Search ​

Vector Embedding Technology ​

Smart Detection ​

User Interface Features ​

Technical Implementation ​

Use Cases ​

🏗️ System Architecture ​

🔄 AI Chat Flow Diagram ​

🛠️ Technology Stack ​

Frontend ​

AI & Machine Learning ​

Backend & Database ​

Maps & Visualization ​

Security & Validation ​

🚀 Getting Started ​

Prerequisites ​

Installation ​

📁 Project Structure ​

🔌 API Endpoints ​

Core Data APIs ​

AI Chat API ​

Analytics APIs ​

🎨 Features Showcase ​

Intelligent Conversational Interface ​

Advanced Data Visualization ​

AI-Powered Similarity Search ​

Comprehensive Filtering System ​

Performance Optimizations ​

🔒 Security Implementation ​

AI Chat Security ​

Data Protection ​

🌟 Amazon Nova Integration Highlights ​

Why Amazon Nova 2 Lite? ​

AI Pipeline Architecture ​

Real-world Impact ​

🚀 Deployment & Production ​

Hosting: Vercel ​

Performance Monitoring ​

� Project Summary ​