Teaching and Data Science
Last Updated: 27, July 2020 at 09:35:49
Overview
Jupyter notebooks combine an interactive read-evaluate-print-loop (REPL) with markdown text (with support for LATEX) to offer a unique solution for people wanting to teach, explore OCaml or perform more data-centric operations (ML models etc.)
From a learning point-of-view, students are more concerned with syntax and concepts rather than tooling (in the beginning). Much like utop the REPL format is great for learning, but with jupyter you can also write notes, explain algorithms etc.
Recommended Workflow
Locally Running an OCaml Jupyter Notebook
The best way to get ocaml-juptyer up and running is through opam and installing jupyter using pip the python package manager. More information on python and pip can be found in their documentation.
With pip and opam installed, the following should represent a typical way of getting started with a jupyter notebook with an OCaml kernel.
$ pip install jupyter # Install jupyter
$ opam install jupyter # Install jupyter opam package
$ opam install jupyter-archimedes # Jupyter-friendly 2D plotting library
$ ocaml-jupyter-opam-genspec
$ jupyter kernelspec install [ --user ] --name ocaml-jupyter "$(opam var share)/jupyter"
Further documentation can be found at in the OCaml Jupyter repository.
Presentation tools
For teachers or presenters you can use the RISE tool to present your code and text in the browser from a running OCaml Jupyer Notebook. It will generate interactive slides from the code and text blocks in your notebook.
Alternatives
Using Docker to run a Notebook
Using the link below, a jupyter notebook can be created inside a docker container and connected to a port so users can begin coding fairly quickly. This involves a second level of indirection in installing docker.
- Limitations - the dockerfiles exposed here are very data-science oriented (i.e. large images with many preinstalled packages) and only go up to OCaml 4.07.
docker run -it -p 8888:8888 akabe/ocaml-jupyter-datascience -v $PWD:/notebooks akabe/ocaml-jupyter-datascience
# Go to http://localhost:8888/?token=<token>
akabe/docker-ocaml-jupyter-datascience