ClimateIQ AI - Climate Intelligence Platform

Inspiration

The Climate Data Gap: While satellites collect terabytes of climate data daily, farmers struggling with crop failures, cities battling air pollution, and communities seeking climate education lack access to this intelligence. ClimateIQ AI transforms satellite data and AI into actionable climate solutions for everyone.

What it does

5 Climate Solutions Powered by Real-Time Data + AI:

🌾 Smart Agriculture (Hybrid ML + GenAI)

Multi-Agent Crop Advisor - Production AI system with 4 specialized agents:

SoilAgent analyzes satellite soil data (SoilGrids 250m resolution: pH, nitrogen, organic carbon)
WeatherAgent predicts climate patterns using OpenWeather API
MarketAgent researches crop prices and profitability (adapts currency: ₹/$/£)
Orchestrator coordinates all insights, resolves conflicts, identifies top 3 crops with phased action plans (immediate/short-term/long-term)

Hybrid Intelligence: Random Forest model (2200+ samples) informs Gemini 2.0 Flash agents for accurate recommendations with confidence scores (60-90 range), reasoning steps, and source citations.

Carbon Footprint Calculator - Gamified accordion UI with live emission meter using IPCC factors and country-specific grid emissions (42 countries, IEA/IPCC 2025).

Solar Irrigation Calculator - NREL PVWatts API calculates solar potential; Gemini agent dynamically searches current government subsidies (PM-KUSUM for India, REAP for USA, etc.).

🗺️ Climate Map Intelligence

6 real-time satellite layers with context-aware scopes:

NASA FIRMS: Thermal anomalies (fires, industrial heat) - 100km radius
OpenWeather: Air quality (AQI), temperature, precipitation - point-based
NREL PVWatts: Solar potential (kWh/m²/day)
Planet API: Vegetation health (NDVI index)
Flood Risk: Rainfall pattern analysis

AI Climate Alert System - Gemini agent analyzes conditions in real-time, generating alerts for:

Heatwaves (>35°C), coldwaves (<5°C)
Poor air quality (AQI 3-5)
Flood risk (>10mm/h rainfall)
Thermal anomalies (>40°C)

5 severity levels (info → advisory → warning → critical → emergency) with actionable safety recommendations.

🏙️ Urban Sustainability

Gemini Vision Waste Scanner - Upload waste images for instant AI classification (plastic, paper, metal, organic, glass, e-waste) with disposal instructions and carbon impact. Privacy-first: no image storage.

Smart Home Tracker - Real consumption tracking (water/electricity) with AI agent searching tariff plans, rebates, and energy-saving tips.

Solar Savings Calculator - 4-step wizard with NREL analysis, ROI calculations, AI-powered subsidy and installer search.

Eco-Advisor - Weather-based sustainability tips + Gemini agent searching city-specific green initiatives (public transport programs, composting, events, rebates).

� Climate Education

Storyblok CMS learning modules with AI-generated quizzes (elegant modal UI, 60% passing threshold) and climate news.

♿ Inclusive Access

Accessibility toolbar with keyboard shortcuts, voice navigation, high contrast themes.

How we built it

Climate Data Stack:

NASA FIRMS (satellite fire detection)
OpenWeather (weather, AQI, forecasts)
NREL PVWatts V8 (solar irradiance calculations)
Planet API (vegetation metadata)
SoilGrids (250m global soil properties)
IEA/IPCC 2025 (emission factors for 42 countries)

AI/ML:

Google Gemini 2.0 Flash Exp (multi-agent coordination, web search, analysis)
Gemini Vision (waste classification)
Custom Random Forest (crop prediction, 2200+ samples)

Platform: Next.js 14, React, TypeScript, Tailwind CSS, Supabase (auth + PostgreSQL with RLS), Mapbox GL JS, Storyblok CMS, shadcn/ui

Challenges we ran into

1. Multi-API Climate Data Integration - Orchestrating 7+ climate APIs with different rate limits, response formats, and coverage areas required intelligent caching and error handling strategies.

2. 100% Real Data Commitment - Eliminated all dummy/fallback data. When APIs fail, users see transparent error messages explaining the issue and alternatives, maintaining trust.

3. Production Multi-Agent AI - Built reliable agent coordination where SoilAgent, WeatherAgent, and MarketAgent provide specialized insights, while Orchestrator resolves conflicts and creates unified recommendations with confidence scoring.

4. Global Climate Context - Auto-detects country from GPS coordinates, adapts currency symbols (₹/$/£), climate data sources (IMD/NOAA/Met Office), and local agricultural boards for any nation.

5. Gemini Response Reliability - Long prompts caused invalid JSON responses. Fixed using Gemini JSON mode (responseMimeType: "application/json") for guaranteed valid output.

Accomplishments that we're proud of

✅ Real-Time Climate Intelligence - 6 satellite data layers covering fires, air quality, vegetation, solar, temperature, and floods

✅ Production AI Agents - 4 specialized Gemini agents coordinating insights with transparency (confidence scores, reasoning steps, source citations)

✅ Hybrid ML+GenAI - Random Forest predictions inform GenAI agents for nuanced agricultural advice

✅ Dynamic Climate Data - AI web search agents find current subsidies/initiatives, eliminating outdated hardcoded data

✅ Climate Safety Alerts - Real-time monitoring generating accurate warnings with expiration timestamps

✅ Privacy-First Design - Waste scanner processes images instantly without storage/tracking

✅ Global Accessibility - 42 countries supported with localized climate data and emission factors

What we learned

Satellite data democratization through user-friendly visualization
Multi-agent AI architecture for complex decision-making with Promise.all parallel execution
Climate API integration at scale (NASA, NREL, Planet, OpenWeather, SoilGrids)
IPCC carbon calculation methodologies and country-specific emission factors
Gemini JSON mode for reliable structured outputs
Accessibility standards (WCAG) for inclusive climate action
Storyblok headless CMS for dynamic educational content

What's next for ClimateIQ AI

🌍 Expand Climate Data: Integrate more satellites (Sentinel-2 for agriculture, MODIS for ocean temperature)

📱 Mobile App: React Native with offline climate data caching

🌐 Multi-language: Translate to Hindi, Spanish, French for 2B+ people

🤝 Community Challenges: Gamified climate action with leaderboards

🔔 Real-time Alerts: Push notifications for climate events in user's area

🏆 Partnerships: Collaborate with climate NGOs and government agencies

Demo Access: Most features work without login. Learning modules & quizzes require authentication.
Test Credentials: Email : test@gmail.com Password : 12345678