Visualizing analytic raster data in the browser still typically depends on backend services that render pre-computed raster tiles (e.g. PNG/JPEG). Aside from being inaccessible — it requires organizations to host and maintain a service — it slows down the iteration cycle for visualization. Every visualization change requires a server round trip to fetch new data.

With this new WebGL-based raster data visualization stack, changing visualization parameters is instant. We can seamlessly animate over time in Zarr data at 60 frames per second, even while applying custom band math.

This browser-based data visualization for COG and Zarr is exposed to Python through Lonboard, the fastest and most capable interactive data visualization tool for Python. This talk will show how to use Lonboard for scalable raster data visualization, and discuss how new tools like async-geotiff enable Python to proxy data to the browser. This allows for seamless, reliable visualization: any COG or Zarr you can access in Python can be visualized. No need to set up a separate tile server.

This talk will give a high-level overview of how these libraries work, and how you can use them in your own projects.