This workshop showcases earthkit as a unified Python pipeline for working with meteorological data -- from heterogeneous sources through to analysis and ML-ready outputs.
The key message: earthkit’s from_source() abstraction means data from CDS, Polytope, local files, S3, FDB, or custom plugins all flow through the same transforms into a consistent output format.
The notebooks are ordered as a learning path. Each builds on the previous, with the thread throughout being: every step gets your data closer to a usable result.
Tutorials
Eight notebooks taking you from loading ERA5 data through to a PyTorch training loop.
Target audiences
EUMETNET-AI workshop participants from national meteorological and hydrological services, and the DestinE community at DUX#5.