At CarbonPlan, we both leverage and contribute to open source software projects that enable our work and the work of the broader climate community. Sometimes that includes participating in community events, like the annual Scientific Computing with Python (SciPy) conference, which members of our team attended last month in Austin, Texas.
The SciPy conference brings together researchers and software developers from a wide range of scientific fields for a week of tutorials, presentations, sprints, and tacos. For our team, highlights from this year’s conference included events related to Pangeo and Xarray along with the Earth, Ocean, Geo, & Atmospheric Science minisymposia, which featured an inspiring lineup of presentations highlighting advances in key data analytics and visualization libraries.
We enjoyed the opportunity to gather with members of the Pangeo community, which focuses on the development of software and infrastructure to enable open, reproducible, and scalable science. During a “Birds of a Feature” discussion session, the Pangeo community reflected on short- and long-term development goals. Ryan Abernathey showcased the Pangeo Forge team’s progress towards crowdsourcing open data in the cloud during the Data Life Cycle minisymposia. We particularly appreciated Julius Busecke’s presentation during the Earth, Ocean, Geo, & Atmospheric Science minisymposia on cloud infrastructure for climate science, which showcased a similar tech stack to the one we used in our recent climate downscaling work.
We were also delighted to see the growing adoption of Xarray, one of the tools in the scientific Python ecosystem that our team both regularly uses and contributes to. During this year’s conference, Anderson taught fundamental Xarray concepts by working through increasingly complex real-world analysis examples in the tutorial Xarray: Friendly, Interactive, and Scalable Scientific Data Analysis . Max presented on behalf of the Generic Mapping Tools community on geospatial analysis and visualization with PyGMT, including visualizing Xarray data objects, and participated in the Xarray sprint, which focused on creating custom indexes and tackling open issues. Joe gave an update on the Xarray project as part of the plenary tools session. Note: While all of the presentations were recorded, some have not been uploaded yet. We’ll add links here when the videos come online.
The keynote on Friday by Peter Kalmus tied several of these themes together, highlighting how open source software accelerates scientific discovery in climate science. Peter featured some of the open source libraries we use the most, like Xarray and Cartopy, and gave examples of how his team uses the open source scientific Python ecosystem to research climate issues ranging from heat waves to coral bleaching.
We returned from the conference energized to continue our work building open tools and resources for robust climate solutions.