diff --git a/docs/_static/geojupyter_cloud_data.png b/docs/_static/geojupyter_cloud_data.png new file mode 100644 index 0000000..13bf7a3 Binary files /dev/null and b/docs/_static/geojupyter_cloud_data.png differ diff --git a/docs/related.md b/docs/related.md new file mode 100644 index 0000000..8c9dae4 --- /dev/null +++ b/docs/related.md @@ -0,0 +1,10 @@ +# Related guides + +Check out these related guides and cookbooks! + +- [NASA Earthdata Cloud Cookbook](https://nasa-openscapes.github.io/earthdata-cloud-cookbook/) - Openscapes cookbook to support NASA Earth Science Team's Migration to the Cloud +- [Scaling Science](https://developmentseed.org/scaling_science/docs/index.html) - Development Seed guide on science applied in industry +- [Zarr Visualization Report](https://github.com/NASA-IMPACT/zarr-visualization-report) - NASA VEDA guide on Zarr visualization approaches +- [Warp resample profiling](https://github.com/developmentseed/warp-resample-profiling) - Development Seed overview of all the Python resampling/reprojection options, built with support from NASA ODSI +- [Cloud-native geospatial formats guide](https://guide.cloudnativegeo.org/) - CNG resource on cloud-native formats, built with support of NASA, CNG, and Development Seed +- [PO.DAAC Cookbook](https://podaac.github.io/tutorials/) - Physical Oceanography Distributed Active Archive Center (PO.DAAC) coobook to make NASA’s ocean, climate, and surface water data universally accessible and meaningful diff --git a/docs/worst-practices/fsspec-caching-defaults.md b/docs/worst-practices/fsspec-caching-defaults.md index 8f164b1..d1f2e87 100644 --- a/docs/worst-practices/fsspec-caching-defaults.md +++ b/docs/worst-practices/fsspec-caching-defaults.md @@ -1,3 +1,7 @@ # Default FSSpec caching arguments FSSpec's default caching behavior is designed for tabular data and performs poorly for datacubes. Recommendations for better better arguments can be found in the [Xarray's tutorial](https://tutorial.xarray.dev/intermediate/remote_data/remote-data.html#remote-access-and-file-caching). + +For example, see the figure below from the [Open Source Geospatial Workflows in the Cloud workshop](https://agu2025.workshops.geojupyter.org/modules/data-in-the-cloud/) showing how much slower it is to open a datacube via `xarray.open_mfdataset()` when using the default readahead fsspec caching mechanism vs. a block cache. For a bonus, you can provide even faster access by virtualizing your data using VirtualiZarr and Icechunk! + +![Figure from Open Source Geospatial Workflows in the Cloud workshop showing data loading performance](../_static/geojupyter_cloud_data.png) diff --git a/mkdocs.yml b/mkdocs.yml index d4dbbc3..a459cef 100644 --- a/mkdocs.yml +++ b/mkdocs.yml @@ -40,8 +40,8 @@ nav: - Xpublish-tiles: "visualization/xpublish-tiles.md" - Rendering layers: - Carbonplan-maps: "visualization/carbonplan-maps.md" - - Deck.gl-raster-layer: "visualization/deck.gl-raster.md" - - Numeric-data-layer: "visualization/zarr-layer.md" + - Deck.gl-raster: "visualization/deck.gl-raster.md" + - Numeric-data-layer: "visualization/numeric-data-layer.md" - Zarr-cesium: "visualization/zarr-cesium.md" - Zarr-gl: "visualization/zarr-gl.md" - Zarr-layer: "visualization/zarr-layer.md" @@ -51,6 +51,7 @@ nav: - API Reference: - Benchmarking: - "api-reference/datacube-benchmark.md" + - Related Guides: "related.md" watch: - packages