AirNet — Real-Time Air Quality Intelligence Platform

Government-grade air quality monitoring, AI-powered policy simulation, and atmospheric analytics for Delhi NCR.

AirNet is a full-stack web application that pulls live air quality data from the WAQI (World Air Quality Index) network, enriches it with Google Gemini AI analysis, stores zone-level readings in a SQLite database, and presents them on an interactive real-time dashboard. A built-in policy simulation engine lets users model the impact of government interventions — such as traffic restrictions and industrial caps — on projected AQI outcomes.

---

Table of Contents

• Overview
• Features
• Tech Stack
• Project Structure
• Getting Started
• API Reference
• AI and Simulation Models
• Database
• Security
• License

---

Overview

System	Description
System 1 — Live Sensor Feed	Syncs real-time AQI, PM2.5, PM10, NO2 data from WAQI for 20 Delhi NCR zones every 30 minutes
System 2 — 24h AQI Forecast	Pulls 2-day hourly forecasts from Open-Meteo and anchors them to live WAQI baselines per zone
System 3 — AI Analysis	Uses Gemini AI to estimate zone-level stress, spike probability, vulnerability, and pollution source
System 4 — Policy Simulation Engine	Deterministic, sub-200ms simulation of government interventions using weighted impact matrices and Gaussian plume dispersion

---

Features

• Interactive NCR Map — Live AQI heatmap of 20 Delhi NCR zones with pollutant drill-down
• Analytics Dashboard — Trend charts, city comparisons, hotspot detection, and 30/60-day historical baselines
• AI Policy Recommendations — Gemini AI generates ranked, source-specific intervention strategies per zone with disruption vs. impact scatter analysis
• Policy Simulation — 5-slider real-time simulation (traffic, industrial, heavy vehicles, dust, construction) with projected AQI drop and recovery time
• Government Policy Generator — AI-assisted policy drafts with Gaussian Plume atmospheric dispersion modelling for quantified impact estimates
• Reports API — Structured API endpoints for programmatic access to all analytics data
• Auto-sync — WAQI data synced every 30 minutes; database auto-pruned to a rolling 60-day window

---

Tech Stack

Layer	Technology
Frontend	Vanilla HTML5, CSS3, JavaScript
Backend	Node.js, Express 5
Database	SQLite3 via Sequelize ORM
AI Engine	Google Gemini API (gemini-2.0-flash / gemini-1.5-flash)
Air Quality Data	WAQI API (World Air Quality Index)
Weather Forecast	Open-Meteo AQI API
Security	Helmet.js, CORS, dotenv

---

Project Structure

airnet/
├── api/                        # REST API route handlers
│   ├── analytics.js            # Trend analysis, city comparison, hotspot detection, zone history
│   ├── policy.js               # AI policy generation, Gaussian plume simulation, recommendations
│   ├── reports.js              # Structured reporting endpoints
│   ├── sensors.js              # Live sensor data & WAQI sync trigger
│   └── simulate.js             # Fast deterministic policy simulation engine (<200ms)
│
├── config/
│   ├── apis.js                 # Centralised API base URL configuration
│   ├── database.js             # Sequelize/SQLite connection setup
│   └── impactWeights.json      # Sector emission reduction weights for simulation
│
├── db/
│   └── schema.sql              # SQLite schema (zone_readings, AtmosphericReadings, Policies, ImpactSimulations)
│
├── models/
│   ├── index.js                # Sequelize model registry & DB init
│   ├── AtmosphericReading.js   # City-level AQI readings model
│   ├── ZoneReading.js          # Zone-level historical readings model
│   ├── Policy.js               # Government policy model
│   └── ImpactSimulation.js     # Policy simulation result model
│
├── scripts/
│   └── generate_historical_policy.js  # Utility: seed historical policy data
│
├── utils/
│   └── waqiService.js          # WAQI sync, Gemini AI analysis, Open-Meteo forecast, DB write, pruning
│
├── api.js                      # Client-side API helper (frontend fetch utilities)
├── api_server.js               # Express API server entry point (port 4000)
├── server.js                   # Static file server entry point (port 3005)
├── app.js                      # Main frontend dashboard application
├── index.html                  # Main dashboard page
├── styles.css                  # Dashboard styles
├── home.html                   # Landing / home page
├── home.js                     # Home page logic
├── home.css                    # Home page styles
├── .env.example                # Environment variable template
└── package.json

---

Getting Started

Prerequisites

• Node.js v18 or higher
• A WAQI API token — Get one free at aqicn.org
• (Optional) A Google Gemini API key — Get one at aistudio.google.com — enables AI zone analysis and policy generation. Without it, the app falls back to a local scoring engine.

1. Clone the Repository

git clone https://github.com/PranavGupta1406/AirNet.git
cd AirNet

2. Install Dependencies

npm install

3. Configure Environment Variables

cp .env.example .env

Open .env and fill in your values:

API_PORT=4000
FRONTEND_PORT=3005

# Google Gemini AI (optional but recommended)
GEMINI_API_KEY=your_gemini_api_key_here

# WAQI Real-Time Air Quality Feed (required)
WAQI_API_KEY=your_waqi_api_key_here
WAQI_CITY_URL=delhi

4. Run the Application

Development (both servers):

npm run dev

Production (frontend only):

npm start

This starts both the static file server on http://localhost:3005 and the REST API server on http://localhost:4000.

5. Open in Browser

Navigate to http://localhost:3005

---

API Reference

The REST API is served on http://localhost:4000/api/v1.

Sensors

Method	Endpoint	Description
GET	/sensors/live	Fetch latest AQI readings for all zones
POST	/sensors/sync	Manually trigger a WAQI data sync

Analytics

Method	Endpoint	Description
GET	/analytics/trends?city_id=delhi&days=7	Hourly AQI trend data
GET	/analytics/compare	Latest AQI for all tracked Indian cities
GET	/analytics/hotspots?threshold=200	Cities exceeding a given AQI threshold
GET	/analytics/zone-trend?zone=Rohini&hours=24	Zone-level trend metrics
GET	/analytics/zone-history?zone=Rohini	60-day historical baseline for simulation

Policy

Method	Endpoint	Description
GET	/policy	List all policies
POST	/policy/suggest	Generate an AI policy from live hotspot data
POST	/policy/simulate	Run Gaussian plume impact simulation for a policy
POST	/policy/recommend	Get AI-ranked zone-specific interventions
GET	/policy/ai-status	Check whether Gemini AI or the local engine is active
PATCH	/policy/:id/status	Update a policy status (Draft / Active / Archived)

Simulation Engine

Method	Endpoint	Description
POST	/simulate/simulate-policy	Real-time deterministic simulation with 5 intervention sliders

Health

Method	Endpoint	Description
GET	/health	API server health check

---

AI and Simulation Models

Gemini AI Integration

When a valid GEMINI_API_KEY is set, Gemini AI is used to:

• Estimate zone-level AQI, stress index, spike probability, and primary pollution source
• Generate specific, legally-grounded government policy directives
• Rank intervention strategies by efficiency (AQI drop per unit disruption)

The app gracefully falls back to a deterministic local scoring engine if the Gemini key is missing or quota is exhausted.

Gaussian Plume Dispersion Model

The policy impact simulation implements the standard Gaussian Plume atmospheric dispersion formula for city-scale AQI forecasting, using empirical sector coefficients calibrated to India CPCB and WHO policy effectiveness studies:

Sector	Emission Reduction Coefficient
Industrial	0.42
Transport	0.38
Energy	0.32
Multi-sector	0.30
Waste / Biomass	0.18

Policy Simulation Engine (System 4)

A fast, deterministic model using a weight matrix across five intervention types:

Intervention	Weight
Traffic Flow Reduction	32%
Industrial Emission Cap	28%
Heavy Vehicle Restriction	18%
Dust Mitigation	12%
Construction Halt	10%

Simulations include wind speed attenuation, 60-day historical AQI blending, trend direction modifiers, and peak volatility buffers, producing results in under 200ms.

---

Database

AirNet uses SQLite (auto-created on first run via Sequelize). The database is not committed to version control.

Table	Purpose
zone_readings	30-minute zone-level AQI readings for 20 NCR zones (auto-pruned to 60 days)
AtmosphericReadings	City-level readings with full AI analysis JSON blob
Policies	AI-generated or manually created policy drafts
ImpactSimulations	Gaussian plume simulation results linked to policies

---

Security

• Never commit your .env file. It is listed in .gitignore.
• Use .env.example as a safe template for collaborators.
• API keys are only read server-side and are never exposed to the frontend.
• Helmet.js and strict CORS rules protect the API server.

---

License

ISC License (c) 2026 Pranav Gupta