Open
Description
We recently decided to remove a logistic regression notebook #82 because it had very high overlap with a Bambi example (https://bambinos.github.io/bambi/main/notebooks/model_comparison.html) and only had the Bambi code.
There is obviously still scope for a logistic regression notebook, but in line with comment by @aloctavodia (#82 (comment)) we should make this example:
- Focus on PyMC code (not rely on Bambi)
- Provide a link to relevant Bambi examples/docs
- Focus on explanation + visualisation
- A suggestion here would be to focus on setting priors and how that can lead to unintuitive prior predictions. (Similar to what @AlexAndorra does in a private PyMC Labs workshop notebook)