I like to use Jupyter for my day-to-day exploratory data analysis tasks. A Jupyter provides interactivity with my code that I explore with by allowing to check output of each code snippet and make changes on the go. A Jupyter combines code, documentation + visualizations in a single document makes it an excellent tool for me to record and keep track of my work.
I like to run my Jupyter on Docker. Docker provides isolation from the underlying system for my development tasks so that, I can easily wipe out system and package dependencies I downloaded for my exploratory tasks later. A Docker container is self-contained and it encapsulates everything needed to run my code. So when I do my experiments with Jupyter on Docker I know exactly what dependencies I need to run my applications on development environment and, later on staging and production environments.
Running Jupyter on a Debian (debian-slim) base image as I use debian Docker environment to run most of my applications.
Contains Dockerfile, docker build image and run container commands to bring up a Jupyter Notebook/Lab with following kernels.
-
Python (default)
-
Typescript (on NodeJS)