OGD4All: A Framework for Accessible Interaction with Geospatial Open Government Data Based on Large Language Models

OGD4All is a newly announced framework leveraging Large Language Models (LLMs) to facilitate citizen interaction with geospatial Open Government Data (OGD). It integrates semantic data retrieval, iterative code generation, and secure execution to generate verifiable multimodal outputs, achieving 98% analytical correctness and 94% recall on a benchmark of 199 questions across 430 City-of-Zurich datasets. This framework is significant for practitioners as it demonstrates a method for reducing hallucination risks and enhancing the reliability of AI-driven public data access, promoting transparency and accountability in open governance.