PyData Global 2020

This folder contains the presentation about decision making with Bayesian methods for PyData Global 2020. The presentation is created from a Jupyter notebook by using RISE and is then served with Binder.

Build slides locally

The libraries required to build the slides locally can be installed from the requirements file:

pip install -r requirements.txt

Use RISE (recommended)

RISE allows to generate the slides from the Jupyter notebook editor itself with a single click:

Use nbconvert

To render slides use jupyter nbconvert foo.ipynb --to slides --post serve

Serve slides for presentation

The slides are served using the environment in arviz_sandbox for convenience and speed, see this post in Jupyter Discourse for a detailed description.

To generate the binder shield with the link to the presentations the following steps should be followed:

• Make sure the notebook metadata has "livereveal": {"autolaunch": true}. If
  you create you presentation as a copy of the English one it will already be
  done.
  • This (as you'll see when editing) generates the presentation automatically
    whenever the notebook is opened. To edit the notebook close the
    presentation and modify the cells contents. More details in
    RISE documentation
• Generate the binder link to run the notebook in the sandbox environment.
  There is a helper page nbgitpuller link generator.
  The Binder tab allows to specify
  • Git Environment Repository URL: https://github.com/arviz-devs/arviz_sandbox
  • Git Content Repository URL: https://github.com/arviz-devs/arviz_misc
  • and the file to be opened
• Create a custom shield from Binder docs