A geospatial notebook that's ready before your coffee
Anyone who has installed GDAL knows the afternoon it costs — and that it breaks again on the next machine, or when a colleague needs the same environment. NotebookHub gives you a persistent workspace with the geospatial stack already in place: open the browser and start reading your rasters.
What's already installed
The geospatial environment starts from the Pangeo community image, the same stack used by research groups working with Earth observation data:
- GeoPandas, Shapely and PyProj for vector work
- rasterio and GDAL for rasters, reprojection and formats
- xarray, dask and zarr for multidimensional data and time series
- The usual suspects: numpy, pandas, matplotlib, scikit-learn
Big files need real disk, not a session
Ephemeral notebooks are a bad fit for geospatial work: a single scene can be gigabytes, and re-uploading it every morning is not a workflow. Your workspace keeps a persistent volume — download once, work for months.
CPU and RAM are reserved for you (Kubernetes Guaranteed QoS), so a heavy reprojection doesn't compete with someone else's job.
And when the analysis has to be seen
The analysis usually ends in a map someone else has to open. GISDeploy, our OGC server, publishes your layers as WMS, WMTS and WFS from the same account — the notebook produces the layer, GISDeploy serves it, and your client opens it in QGIS or in the browser.
Both products are operated by the same team, in Chile, with support in English and Spanish.
Frequently asked questions
Can I install extra packages?
Yes. conda and pip work normally and the installs persist, because the disk is persistent. If your team needs a specific image, write to us and we'll prepare it.
How much data can I keep?
From 10 GB on the trial to 200 GB on the Team plan, on a persistent volume. If you need more, tell us what you work with and we'll size it.
Does it work with PostGIS?
Yes — it's a normal Python environment, so psycopg and GeoPandas connect to your PostGIS as usual, whether it's yours or ours.
Can I publish the results as a map service?
Yes, with GISDeploy (WMS/WMTS/WFS), operated by the same team. Ask us and we'll connect both accounts.
Try it with your own data
No credit card. Upload a raster and see how it feels.