// the find
OpenGeoscience/geonotebook
A Jupyter notebook extension for geospatial visualization and analysis
A Jupyter notebook extension from Kitware and NASA Ames that embeds interactive geospatial maps (via GeoJS) directly in notebooks, with a built-in tile server backed by Mapnik and GDAL. Aimed at scientists doing raster/vector analysis who want to stay in a notebook workflow rather than switching to QGIS or a web GIS. Last commit was January 2019, so it is effectively abandoned.
The JSON-RPC bridge between the Python kernel and the JavaScript map is a clean architectural choice — notebook cells can drive map state programmatically without any callback spaghetti. Pluggable tile server backends (built-in ktile/Mapnik or external GeoServer) means you are not locked into one rendering path. The VRT wrapper support is genuinely useful for working with large multi-band rasters without loading them into memory. Test coverage is reasonable for a research project — unit tests plus integration notebooks.
Dead project. Python 2.7 references in tox.ini, Jupyter notebook extension APIs it targets were deprecated years ago and are incompatible with JupyterLab 4.x — you would spend more time fighting extension loading than doing any actual analysis. The GDAL >= 2.1 and mapnik >= 3.1 system dependency chain is painful to install on anything modern, and the Docker container has not been touched since 2018 so it probably does not build. No async support anywhere in the kernel communication layer, so large raster operations block the kernel. ipyleaflet or lonboard do everything this tried to do, are actively maintained, and integrate with the current Jupyter ecosystem.