Add notebook: case study with item response theory

[ltoniazzi]

Description

This notebook performs data analysis of NBA foul calls with Item Response Theory. It is based on a blog post of @AustinRochford. The emphasis is on these steps:

1. we have a real world complex problem with a sizeable dataset,
2. so pick an “ambitious’’ model,
3. show that is easy to instantiate it and solve it with PyMC3,
4. and finally discover something by exploring the posteriors.

Reasons for adding to case_studies:

• The Rasch model does not appear in the current case studies.
• Good balance of complexity of the model versus ease of PyMC3 implementation.
• It leaves the reader with a suggested extension of the model to experiment further.
• Related notebooks from @AustinRochford allow readers to see this model from many different viewpoints (for example how it compares to simpler or more complex versions) on the same kind of data.

Please be aware that I am a beginner contributor.

Cheers, Lorenzo

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[chiral-carbon]

Hi, thanks a lot for your PR! I have left 2 comments pertaining to the style used in example notebooks as pointed out in this wiki.

[ericmjl]

@OriolAbril I've personally reviewed the notebook with @ltoniazzi before, but I think it'd be good to have further comments from you and the rest of the team, at your convenience 😄.

[OriolAbril]

@OriolAbril I've personally reviewed the notebook with @ltoniazzi before, but I think it'd be good to have further comments from you and the rest of the team, at your convenience smile.

👌 I probably won't have time until next week for a proper review, but if someone else can review and merge please do so, a review from me should clearly not be a blocker for the PR to be merged

[fonnesbeck]

This is a terrific addition to the repository. Thanks @ltoniazzi !

[ericmjl]

Thanks for the work, @ltoniazzi!

[ltoniazzi]

Thanks everyone, glad to contribute!