Climate AI Hive – Goal 11 AI Agent Platform for Sustainable Urban Development

Climate AI Hive is an AI-powered urban climate resilience platform built on IBM WatsonX and IBM Cloud, using IBM’s 🐝 Bee AI agentic framework.
It combines UN Sustainable Development Goal 11 (Sustainable Cities & Communities) data, IPCC climate science, and specialized BeeAI agents, powered by IBM Granite 3.3.

In just minutes, it turns climate change and UN data into city-specific action plans with projections, recommendations, and SDG 11 compliance metrics — empowering cities to adapt faster and smarter.

A project made for IBM TechXchange 2025 Pre-conference watsonx Hackathon

🌍 Problem Statement

Cities face an urgent need to adapt to climate change with limited resources.
Today:

• Climate data is fragmented and hard to use.
• Assessments (e.g., flood risk) take months to complete.
• Projections (e.g., heatwaves) are disconnected from urban planning.
• Thousands of proven climate solutions exist but remain invisible to most cities.
• International funding opportunities are missed due to lack of proper SDG alignment documentation.

The result: cities react to disasters instead of preventing them. Resources are misallocated, proven solutions stay undiscovered, and funding is lost.

💡 Solution

Hive.ai – UN & IPCC Expertise in Your Pocket with BeeAI Agents on IBM WatsonX 🐝

Three specialized AI agents collaborate to produce an integrated climate resilience report:

• BeeAI ClimateAnalyst – Analyzes historical and projected climate data (temperature, precipitation, air quality, flood risk), factoring in model uncertainties and biases.
• BeeAI UrbanAdvisor – Retrieves official UN SDG data via the SDG API, maps infrastructure vulnerabilities and demographics, and generates locally tailored recommendations in mobility, green spaces, energy, waste, and citizen engagement.
• BeeAI SDG11Validator – Evaluates proposals against UN SDG 11 targets, assigns alignment scores, and suggests improvements to maximize compliance and funding eligibility.

Hive.ai delivers:

• Hyperlocal climate impact projections
• Prioritized action lists with budgets and timelines
• Global case studies adapted to local contexts
• SDG 11 compliance metrics and funding guidance

🚀 Key Features

• Climate Impact Analysis – Projections, trends, and vulnerability maps
• Sustainable Recommendations – Context-specific, actionable strategies
• UN Project Discovery – Relevant initiatives and funding sources
• SDG 11 Validation – Measured and improvable alignment scores
• Multi-provider AI – IBM WatsonX and Groq support

🎥 Climate AI Hive Demo

Watch a quick demonstration showing how Hive.ai generates local climate action plans in minutes — including projections, tailored recommendations, and SDG 11 alignment metrics:

🛠️ Installation

Prerequisites

• Python 3.8+
• Node.js 18+ (for frontend)
• IBM WatsonX account (optional, but recommended)

Backend

cd backend
pip install -r requirements.txt
cp env.example .env  # Copy and edit with your values
uvicorn api:app --reload

Frontend

cd the-hive
npm install
npm run dev

🤖 AI Provider Usage

You can configure the default provider and models in your .env file:

# === Watsonx (IBM) ===
WATSONX_API_URL=https://eu-de.ml.cloud.ibm.com
WATSONX_API_KEY=your_watsonx_api_key
WATSONX_PROJECT_ID=your_project_id

# === OpenAI (optional) ===
OPENAI_API_KEY=your_openai_api_key

# === Default Model ===
MODEL_NAME=watsonx:ibm/granite-4-h-small

💡 You can change MODEL_NAME to another model supported by BeeAI, such as: watsonx:ibm/granite-3-3-8b-instruct or openai:gpt-4.1-mini, depending on your needs.

📡 API Endpoints

POST /climate-impact

Analyze climate change impact on a city.

POST /recommendations

Get sustainability recommendations for a city.

POST /un-projects

List relevant UN projects for a city.

POST /sdg11-validation

Validate a proposal’s alignment with SDG 11.

🌟 Specific Use Case

A resilience officer in Phoenix enters their city into Hive.ai. Within 5 minutes, they receive:

• Heat island vulnerability maps with demographic overlays
• Prioritized interventions (urban forests, cooling centers, reflective surfaces)
• Medellín green corridor success case study
• Budget estimates and financing options
• SDG 11 alignment scores
• Implementation timeline with measurable indicators

🧠 Team

We are a multidisciplinary team combining expertise in AI agent architecture, data science, and UX design:

• Paul – co-founder of Inclusive Brains and AI Agent Architect at Wavestone, working with IBM France on Quantum Machine Learning for neuroscience data,
• Tristan – Agent Engineer & Data Scientist at Wavestone, designing and building Hive.ai’s specialized agent tools.
• Louise – AI Engineer & Data Scientist at Wavestone, leading data processing and analysis while crafting intuitive, actionable user experiences.
• Mentor: Olivier Oullier – Neuroscientist & co-founder of Inclusive Brains, guiding UX/UI design and AI strategy.

📝 License

MIT License.