Join us at the Pangeo ML working group meeting at noon ET on Monday, November 07, 2022 for a presentation by Wei Ji Leong (Byrd Polar and Climate Research Center) on "zen3geo: Guiding Earth Observation data on its path to enlightenment”, followed by an opportunity to share individual updates, and concluding with a discussion of agenda items.
When - November 7, 2022 12 PM ET
Where - Launch meeting - zoom
Agenda / Minutes - Meeting-notes-Machine-Learning-WG
Presentation Title - zen3geo: Guiding Earth Observation data on its path to enlightenment
Presentation Abstract -
When spatial data meets machine learning, how do we unite the two while keeping them separate? With torchdata, there is now a way to favour composition over inheritance, chaining together DataPipes that individually, excels at one function, but combined, works to process your custom dataset iteratively in a streaming fashion. Is your spatial data special? Only if you keep your metadata intact, and what better way to store those via xarray and/or geopandas! As a lightweight library, zen3geo is designed for the cloud-native geospatial era, acting as the glue between various geospatial libraries in the Pangeo ecosystem. From reading Spatiotemporal Asset Catalog (STAC) items, to rasterizing vector geometries with datashader and slicing multi-dimensional data with xbatcher, mix-and-match any component or custom transformation as you please. Begin your journey at https://zen3geo.readthedocs.io/en/v0.5.0/walkthrough.html and unlock new possibilities in Machine Learning for Earth Observation and beyond!
Presenter - Wei Ji Leong (Byrd Polar and Climate Research Center)
Will this be recorded? - No