I recently released version 2.2.0 of Zarr, a Python package for numerical data storage. One of the new features available with Zarr is the ability to store and retrieve data directly from cloud object storage systems like Amazon S3 or Google Cloud Storage (GCS). One of the coolest things about working on Zarr is that although I work on genomics, I get to meet people from a number of different scientific disciplines, with similar issues and interests in being able to run interactive analyses over large numerical datasets. In particular, there is some really exciting work going on within the geoscience community around the Pangeo project, which is working to enable better use of cloud infrastructure for ocean, atmosphere and climate data science. The Pangeo project has put together some very nice demos of using Zarr in combination with other packages like Dask and Xarray to run interactive analyses on large datasets via Google cloud. Off the back of that work, I was invited to give a webinar as part of the ESIP tech dive series. In the webinar I tried to give an overview the main architectural elements of Zarr, with some worked examples. Hopefully it’s useful, here it is on YouTube: