Data scientists often switch between Jupyter notebooks and desktop applications to view visualizations. This context switching breaks workflow. A new approach runs a full Linux desktop environment directly inside JupyterLab, eliminating file exports and application switching.

The solution builds on 2i2c's NASA QGIS Desktop repository. It uses the Jupyter minimal notebook image as a base, allowing customization with Pangeo images, Python packages, and Linux applications. This creates a portable, cloud-ready environment.

Developers can run this image locally or on any Jupyter Hub that supports user-supplied images. It streamlines the data exploration pipeline by keeping analysis and visualization in one integrated workspace, reducing friction for teams working with complex geospatial or scientific data.