Oahe Data - Guthrie Ducheneaux & Jeff Lumpkin - September 3, 2026

Guthrie Ducheneaux and Jeff Lumpkin of Oahe Data presented their work building analytical infrastructure, data visualizations, and AI-ready data products. They developed a 'data refinery' that rapidly shapes and standardizes data, which they applied to public datasets, though the process could apply to any data. Raw files move through stages of ingestion, mapping, and validation to produce standardized bundles with unique identifiers, chain-of-custody tracking back to the original source row, and tamper-evident hashes. Data are stored as parquet files and queries with DuckDB for fast extraction and transformation, while a defined set of rules govern how AI can join, format, and query the data, and restricts the AI from using outside information or estimation in order to keep results reliable. This infrastructure powers their Community Needs Assessment, which draws on thousands of distinct, longitudinal measures and can be viewed at the reservation, hospital, or county levels. It includes options to view census tracts, towns/cities, or include overlays with congressional districts. Users can explore the data through maps, a tract explorer, and a graph explorer, then apply AI, which is guided by set query rules, to generate text summaries of adjustable length. Data products and analyses can be saved to a library and assembled into a full report through an automated process. Their goal for these tools is to deliver actionable intelligence faster and make advanced analytics more accessible.