Project Atlantis

A Scalable STAC/Zarr Pipeline for ML-Ready Multi-Source Flood Inundation Observations

Project Atlantis builds a reproducible pipeline that gathers and harmonizes satellite flood‑inundation data into a unified, ML‑ready STAC/Zarr archive. It uses an open‑source Python geospatial stack with Docker and a custom CLI to automate ingestion, standardization, and end‑to‑end ML validation. By transforming fragmented Earth observations into a standardized, high-fidelity data foundation, this project enables the development of more accurate predictive models that are critical for protecting vulnerable communities and mitigating the economic impact of extreme flood events.