Democratizing Climate Analytics: A Local-First Approach using DuckDB, H3, and WASM

Climate and Earth Observation datasets are growing rapidly, yet the infrastructure commonly used to process them can introduce significant overhead and slow analytical iteration. This abstract presents a local-first spatial analytics workflow; validated through operational deployments in Turkey, Madagascar, and the UK; built around DuckDB, an embedded analytical database engine, together with its Spatial and H3 hexagonal indexing extensions. The approach enables reproducible spatial SQL over efficient columnar file formats (Apache Parquet), supporting fast spatiotemporal aggregation, multi-resolution summaries, and boundary integration without heavyweight GIS server infrastructure. Results can be published as a portable database served via lightweight APIs, or explored entirely in-browser using WebAssembly. This work contributes to UN SDG 13 (Climate Action) and SDG 2 (Zero Hunger) by enabling accessible, low-cost climate monitoring for drought early warning and food security.