This post is a follow-up to an earlier post about starting an online analytics portfolio, focusing on R. In that post, I walk though how to get markdown from an R script and share the static results with the internets.
A better way, with Jupyter Notebooks (
Been reading more about jupyter notebooks (
- They are not limited to python kernels – they can also run R.
nbconvertcan convert them to a number of formats, including markdown and html.
- Several places will
nbconvert-them to html for you, including https://nbviewer.jupyter.org/ and github, gists, etc.
- Under the hood, they are
json. They save the results from when their code cells are run. Therefore,
nbconvertdoes not need to execute the code cells. This means that notebooks that run specialized kernels or that need access to specialized environments such as ones from Amazon Sagemaker can be rendered using
- jupyter notebooks can, obviously, be opened and run by jupyter, as long as jupyter has access to a compatible kernel and environment.
- Several sites offer free cloud-based JupyterHub, which can let other people view and run jupyter notebooks, as long as the rendering service has a compatible kernel. These include but are not limited to: