Add causal product incrementality analysis with multivariate interrupted time series models

[drbenvincent]

This PR will add new functionality to pymc-marketing. Namely it will allow causal analysis of product incrementality/cannibalization.

In this initial PR we will focus on:
a) adding docs to introduce the idea behind incrementality/cannibalization analysis
b) in terms of modeling we will start with a multivariate interrupted time series model. This will include a notebook to illustrate its use, as well as docs in the guide section to convey the basic workings.

Later PR's will add additional modeling options, such as nested discrete choice models.

TODO

• Finish "Introduction to product incrementality" docs page
• Finish MV-ITS model explanation
• Add tests
• Finish how-to notebook demonstrating use
  • Add module code
  • Add Scenario 3 where we operate on product level data
  • Add scenario where market is not saturated + update relevant docs + tests
• Add a QuickStart on the Getting Started docs page
  • add python code block for QuickStart
• Rename module to consumer-choice. Update docs and imports and re-run notebooks etc.
• Fix typo "isoluate" -> "isolate"
• Update docs homepage and README.md

Known limitations

At the moment the MVITS model is very simple. It only has an intercept term, embodying the assumption that sales are relatively static over time. Clearly there are many ways to expand this model, so the proposal is that we get this in as a first attempt and stepwise add features over time. We can be lead by feature requests (if there are any), but this establishes an MVP which can be built upon.

Checklist

• Checked that the pre-commit linting/style checks pass
• Included tests that prove the fix is effective or that the new feature works
• Added necessary documentation (docstrings and/or example notebooks)
• If you are a pro: each commit corresponds to a relevant logical change

Modules affected

This PR will create a new module and will not affect the MMM or CLV modules.

Type of change

• New feature / enhancement
• Bug fix
• Documentation
• Maintenance
• Other (please specify):

---

📚 Documentation preview 📚: https://pymc-marketing--803.org.readthedocs.build/en/803/

[drbenvincent]

Everything is passing now, except codecov

[wd60622]

All looks really clean @drbenvincent Nice work!
I will take a look later and can do those edits to docstrings, tests, etc when I get the chance

[drbenvincent]

@wd60622 All tests pass again now - aside from code coverage. Am I right in understanding that you might do some work on this PR? I probably need to put this aside until late next week until things free up a bit.

[wd60622]

@wd60622 All tests pass again now - aside from code coverage. Am I right in understanding that you might do some work on this PR? I probably need to put this aside until late next week until things free up a bit.

Nice. Yes, I take a look today