Jupyter is a useful environment for data exploration. Despite the many excellent visualization packages, it's often more expedient to use a purpose-built application to see results. Typically, results are exported to a file, which is downloaded and imported into the application. The extra steps require context switching from an online application to an application running in on your computer.
Running a Linux desktop in JupyterLab eliminates the need to move data between environments and reduces context switching. Based on 2i2C's NASA QGIS Desktop repository, this repository uses the Jupyter minimal notebook image as the base image. This is a basic image that can be customized by using base images, such as Pangeo images, adding domain-specific Python packages, and installing any Linux applications.
The image can be run locally or on any Jupyter Hub that supports running user-supplied images. Go build something.
https://github.com/spara-earthscope/earthscope-desktop-image
If you're not familiar with 2i2c, check them out, starting with their 2026 strategy update.
Top comments (0)