Multi-Agent Methane Data Query System based on GenAI and DGGS

Objective

Develop a scalable geospatial intelligence framework that integrates Discrete Global Grid Systems (DGGS) and generative AI for automated extraction, normalization, and spatial indexing of environmental data from both structured and unstructured sources, demonstrated through a methane emissions use case.

Description

The Multi-Agent Methane Data Query System is an innovative geospatial intelligence framework that revolutionizes how environmental data is accessed and analyzed through natural language interactions. This system enables users to ask complex questions about methane emissions data using everyday language, such as “What’s the total methane emission from oil and gas in Alberta in 2018?” or “Compare methane emissions between different basins over time.” The system automatically processes these queries through a sophisticated multi-agent architecture that understands user intent, resolves geographic locations, and generates meaningful insights from vast environmental datasets.

The framework demonstrates the powerful integration of generative artificial intelligence with Discrete Global Grid Systems (DGGS) for automated environmental data extraction and spatial analysis. By combining natural language processing capabilities with advanced geospatial indexing techniques, the system provides researchers, policymakers, and environmental scientists with an intuitive interface to explore methane emissions data across different geographic regions, time periods, and industrial sectors. This approach eliminates the need for complex database queries or specialized geospatial software, making critical environmental data more accessible and enabling faster, more informed decision-making for climate change mitigation efforts.

Project Details

Collaborator(s): Dr. Erin Li and Dr. Steve Liang

Date: 2025-ongoing