EDTC WEBINAR

From Workshop to Query: Building a Community-Grounded Flood Knowledge Graph

Thursday, November 19, 2026 · 10:00am MT / 12:00pm ET

Virtual · Free

About the session

This talk describes how NOAA's Weather Program Office Social Science Program built a flood knowledge graph starting from community knowledge rather than formal ontology engineering. A multidisciplinary workshop with the Natural Hazards Center produced conceptual models of flood systems. AI-assisted workflows then turned those models into a unified ontology. The ontology was then grounded against NOAA, FEMA, and Census datasets and a natural-language interface in which an LLM writes the SPARQL queries was added.

Testing several LLMs showed that most query failures came from gaps in the schema, not from the models. That suggests the quality of the semantic layer limits AI-assisted access more than the model does. Jonathon will share what worked, where a workshop-derived ontology falls short, and how the approach could extend to other hazards.

Your presenter

Jonathon Mote

Social Science Program, NOAA Weather Program Office

Jonathon Mote is a contractor supporting the Social Science Program in NOAA's Weather Program Office (WPO), where his work increasingly focuses on knowledge management and data infrastructure. A self-described enthusiast of knowledge graphs and large language models, he has been exploring how community-built ontologies, semantic web tools, and AI can make weather and hazard data easier to integrate and query. Before joining WPO, he was Director of Research Analytics at Digital Science and held faculty positions at the University of Maryland, Southern Illinois University, and George Washington University. He holds a PhD in Sociology from the University of Pennsylvania.

Register

Powered by Semantic Arts