FOSS4G 2022 general tracks

The SpatioTemporal Asset Catalog (STAC) specification is a common language for describing geospatial information that is flexible enough to extend across domains and use cases. In this talk, we walk through best practices for building STAC catalogs and using STAC extensions, using real world examples. These best practices are informed by documentation, conversations with STAC contributors, and discussions within the wider community. We survey the ecosystem of open-source STAC software, which includes libraries and tools written in Python, Node.js, and more. We show examples of reading, modifying, and writing STAC catalogs with a selection of software, including PySTAC and stactools, and we show which metadata to include in your STAC objects to ensure interoperability with powerful tools like xarray and pandas. Whether you are new to the STAC ecosystem or an experienced contributor, this talk will provide you with the context and tools you need to build your best STAC!

As a part of its AI4Earth initiative, Microsoft has created a Planetary Computer (PC) for hosting and processing open geospatial data. In addition to publishing a wide range of datasets, including Sentinel-2, MODIS, and more, the PC provides a powerful API and compute system based on open-source geospatial tools and using STAC metadata for data query, discovery, and access. In this talk, we present the latest in open geospatial data access, discovery, processing, and visualization using a variety of datasets from the Planetary Computer. We demonstrate use of the odc-stac package, which leverages the power of the OpenDataCube computing platform without the need for a database backend, and how odc-stac can load, mosaic, and transform geospatial assets into xarray datasets. We dive into other data interoperability tasks, including scaling processing with Dask and leveraging a variety of cloud-native formats. Along the way, we provide recommendations for data providers and curators on how to ensure their data can be used in a rich, interoperable way by the latest in geospatial processing tools.