GeoCAT-Applications: Building an Open Source Community Resource for Geoscientific Applications in Python

Since the NSF National Center for Atmospheric Research (NCAR) announced its pivot-to-Python, away from the NCAR Command Language (NCL), in 2019, the Geoscience Community Analysis Toolkit (GeoCAT) team has worked to support scientists making the transition between coding languages through tutorials, utility functions (in GeoCAT-comp and GeoCAT-viz), and by developing a one-to-one library demonstrating how to reproduce NCL visualizations in Python (GeoCAT-examples). To demonstrate how to combine different packages to achieve the functionality requested by our user community as they transition to Python, the GeoCAT team began development on a new resource called GeoCAT-applications (https://ncar.github.io/geocat-applications/). GeoCAT-applications is designed to serve as a lightweight reference demonstrating individual components of geoscience workflows in Python. Python is a widely used programming language for data analysis and visualization within the geosciences because it is free, open-source, and relatively easy to learn. Python allows scientists to write working code quickly with access to a wide array of toolkits. However, the seemingly endless and growing pool of packages, each with their own documentation and examples, can be overwhelming, especially to a new Python coder who may not know what to use or where to look. Several groups have worked to address aspects of this challenge. The Pangeo community works collaboratively to develop software and infrastructure to enable Big Data geoscience research and defines and supports an “ecosystem” of interconnected Python packages useful to geoscientists. Project Pythia, a community resource for learning how to analyze geosciences data using the Scientific Python Ecosystem, maintains a library of end-to-end workflows called Cookbooks. However, users transitioning to Python still missed the one-stop-shop, at-a-glance view that the NCL applications page (https://www.ncl.ucar.edu/Applications/) provided. GeoCAT-applications is a unique quick reference guide with grab-and-go, bite-sized coding solutions for demonstrating components of a scientific workflow. GeoCAT-applications has three main components: Python pages that do not presume NCL familiarity, NCL-to-Python entries that document the difference between NCL and Python functionality, and NCL receipts for more detailed comparison testing. This structure allows us to both support NCL users transitioning to Python and provide a more generalized resource for the broader geoscience community.