A Python Workshop To Efficiently Access, Deliver, Visualize, and Analyze Big Environmental Data in the Cloud

This virtual, hands-on Python workshop, presented by Ryan Paul Lafler, provides an applied introduction to accessing, processing, visualizing, and analyzing large environmental and weather datasets directly in the cloud. Designed for Python programmers, data scientists, environmental and climate scientists, researchers, students, and applied practitioners, the workshop focuses on efficient, scalable strategies for working with petabyte-scale data stored in Amazon S3, Google Cloud Storage (GCS), and Microsoft Azure, without downloading entire repositories or duplicating data locally.

Previously presented at the American Meteorological Society (AMS), this workshop gives attendees the practical skills and working knowledge needed to work effectively with large environmental datasets in cloud environments using modern big-data processing and computation strategies. Participants learn how to move beyond local file-based workflows and adopt cloud-native analysis patterns that scale from personal systems to larger computing environments.

Attendees will build Python data pipelines capable of selectively accessing, processing, and retrieving only the environmental information they need, even when datasets span petabytes of open climate and meteorological data. Using modern open-source Python libraries such as Dask for scalable computation, Xarray for multidimensional spatiotemporal data, and fsspec for cloud object storage access, participants will learn how to efficiently process, visualize, and analyze large climatological and meteorological datasets directly from the cloud.

Key topics covered include:

• Connecting to cloud object storage (Amazon S3, GCS, Azure) using Python and fsspec

• Working with real-world datasets, including the Coupled Model Intercomparison Project (CMIP6) and NOAA’s Real-Time Mesoscale Analysis (RTMA)

• Understanding environmental data formats: NetCDF, GRIB2, HDF5, and Zarr

• Efficient data access patterns, including lazy loading, chunking, and graph-based execution

• Building scalable data pipelines with Dask and Xarray

• Extracting time series and performing spatiotemporal analysis

• Visualizing environmental data using Matplotlib, Cartopy, and Rasterio

All participants will receive a Certificate of Completion, the non-redistributable PDF slides, a documented Jupyter Notebook, and cloud data access examples to support continued practice, skill-building, and independent learning after the workshop.