CISL Seminar: GeoCAT and Creating Python Software for Big Data and HPC

Abstract

The Geoscience Community Analysis Toolkit, GeoCAT, was created as part of the National Center for Atmospheric Research’s “Pivot to Python” initiative. The GeoCAT team’s mission revolves around creating an open-development, scalable Python based geoscience computation and visualization toolkit. This mission is realized through our team's implementation of a continuous integration, open-source software development structure to encourage community involvement at every stage of the project lifecycle -- from requirements and specification to testing and deployment.

As a part of NCAR’s Science at a Scale initiative, the GeoCAT team was asked to demonstrate some computational routines using data from the NCAR Community Earth System Model (CESM) Large Ensemble (LENS) hosted on Amazon Web Services (AWS). The computational functions created were designed to be compatible with dask and xarray, and were demonstrated on NCAR’s JupyterHub on Casper.

This seminar will cover an overview of GeoCAT, the development process of computational routines for the Science at a Scale initiative, and what we learned about developing in Python for big data and high performance computing along the way.

Bio

Anissa Zacharias is a software engineer at the National Center for Atmospheric Research (NCAR). She is a member of the GeoCAT team and works on computational and visualization tools for geoscience applications. She has a MS in computational science from Georgia Tech and a BS in computational mathematics from Embry Riddle Aeronautical University.

For more information, contact taysia@ucar.edu.