Jupyter Notebooks Linked to Other Applications

[psychemedia]

Data sources used for scripts running in Jupyter notebooks may often be accessed via a database. The practical task of setting up databases, seeding them with data, and connecting to them from a notebook can be an issue.

One way round this is to define simple recipes for running a set of linked applications, eg using Docker Compose. For running workshops, providing a set-up that provides a participant with a complete environment - Jupyter notebook + DBMS + seeded data - can allow folk to start working on the data without having to worry about the initial sys admin/database management worries. Running single Docker containers is easy enough on the desktop using something like Kitematic, though it still requires issuing a command line command (from the Kitematic UI) to run multiple linked containers via Docker Compose.

An alternative to Docker that may work in simple workshop settings could be to create custom Raspberry Pi SD card images, with software and data preinstalled, that just free runs Jupyter on a known port. The downside is the the RPi is not that powerful, though it should be okay for simple workshops?