ACCESS 2019 Kickoff OCT 29 2020 ACCESS 2019

ACCESS 2019 Kickoff OCT 29 / 2020 ACCESS 2019 KICKOFF PANGEO-ML SLIDE 01/07

● Pangeo is a community promoting open, reproducible, and scalable science. ● Pangeo is an integrated ecosystem of open source software tools. ● Pangeo is a community platform for Big Data Geoscience. ACCESS 2019 KICKOFF PANGEO-ML SLIDE 02/07

ACCESS 2019 KICKOFF PANGEO-ML SLIDE 03/07

Analysis Ready Data Stored and cataloged on globally-available distributed storage (e. g. S 3, GCS) Jupyter for interactive access on remote systems Distributed storage Cloud/HPC Parallel computing system built on top of Kubernetes or HPC. Xarray provides data structures and intuitive interface for interacting with datasets Dask tells the nodes what to do. ACCESS 2019 KICKOFF PANGEO-ML SLIDE 04/07

Team Joe Hamman Carbon. Plan/NCAR ACCESS 2019 KICKOFF Ryan Abernathey Columbia U. David Hoese U. Wisconsin PANGEO-ML James Bednar Anaconda Inc. Tom Augspurger Anaconda Inc. SLIDE 05/07

Pangeo-ML HIGHLIGHTS ● ● Improve interoperability of the scientific Python ecosystem (Pangeo, Pytroll, ML Libraries) Development of new software interfaces connecting Xarray and ML libraries (e. g. xbatcher) Expansion of the Earth. ML open-source documentation platform Continued improvement, maintenance, and support of Pangeo/Pytroll software ACCESS 2019 KICKOFF One of the central goals of the Pangeo. ML effort is to enable seamless transitions between exploratory data analysis and machine learning applications. PANGEO-ML SLIDE 06/07

Pangeo-ML Monthly Meeting Focus (to date): ● ML workflows on geospatial data ● Pipeline tools that work with Pangeo ● Infrastructure needs Open to anyone. More details: https: //pangeo. io/meeting-notes. html ACCESS 2019 KICKOFF PANGEO-ML SLIDE 07/07