In this session, participants learn how to access and process Earth Observation data in a cloud-native workflow using STAC APIs, xarray, stackstac, and Dask. Using a post-flood change analysis in Valencia as a case study, they build before/after composites from Sentinel-2 (optical) and Sentinel-1 RTC (SAR) and apply unsupervised change detection (clustering in spectral and radar feature space) to identify multiple change patterns without task-specific labels. Hands-on exercises show how to compare optical and SAR change signals, interpret cluster-based change modes, and extend the workflow toward AI-ready multimodal analysis.
For any questions, contact @albughdadim.