DOCS Updating the Rugby docs

[springcoil]

I've updated the Hierarchical Rugby model in the Documentation.

I've added - more data from more years, since before it only had 2014.

I've updated a few things in terms of the explanations, added a few more visualizations/ exploratory data analysis and added a few links to other resources.

I'd appreciate some proof reading, and for someone to verify if they run correctly.

[springcoil]

Anyone know why this build is breaking? @twiecki @junpenglao ?

[ColCarroll]

It looks like travis randomly killed everything... I restarted all builds

[ColCarroll]

Oh, you might need to rebase off master: we pinned Theano to 0.9.0 for some reasons (#2728), but it looks like it is trying to install Theano 1.0.1 right now.

Notebook diffs are hard to see what changed! Generally there are a few edits I would make on the introduction. Feel free to ignore (especially ones that aren't associated with this change), and I can make a bunch of nits myself in a future PR.

• First sentence should be something like

I came across the following blog post on Daniel Weitzenfeld's blog, based on the work of Baio and Blangiardo.

Note that I changed the link for the Baio and Blangiardo paper, since the old one is dead.

• Remove the bullet points in the introduction, since most of them are not referred to later. The first (about which countries are in the Six Nations) is useful to know, but should be included in the previous sentence:

Since I am a rugby fan I decide to apply the results of the paper to the Six Nations Championship, which is a competition between Italy, Ireland, Scotland, England, France and Wales.

• The last sentence of the introduction should be updated for the new data

• The "Motivation" section might also be reworked. There is some capitalization which should not be there, and a stray close parens. I would drop the TODO as well.

• You might reorder the data_csv_201X section to be chronological. You could also add a "year" column, and dump the csv in "pymc3/examples/data/rugby.csv", then use df_all = pd.read_csv(pm.get_data('rugby.csv'))

[springcoil]

I made those changes you suggested Colin.

[springcoil]

If this builds correctly we can merge it I think. There's a few nitpicks etc, but I could fix them or @ColCarroll at some other point.

[ColCarroll]

This is really good - I like that we have a new data set! Feel free to merge whenever - you did not write any code that is run in the test suite, so travis is not going to say anything useful. (I am not merging right now because I am interested to see if it passes, or if travis is still unhappy for other reasons)

[springcoil]

I want to see that too. I think it's awesome we've another data set?

…

On 29 Dec 2017 17:22, "Colin" ***@***.***> wrote: This is really good - I like that we have a new data set! Feel free to merge whenever - you did not write any code that is run in the test suite, so travis is not going to say anything useful. (I am not merging right now because I am interested to see *if* it passes, or if travis is still unhappy for other reasons) — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#2766 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AA8DiC5EdwX51m8eh58lQYRfKgOwyooMks5tFRGugaJpZM4RPC7D> .

[springcoil]

Seems to still be failing....

…

On 29 Dec 2017 17:39, "Peadar Coyle" ***@***.***> wrote: I want to see that too. I think it's awesome we've another data set? On 29 Dec 2017 17:22, "Colin" ***@***.***> wrote: > This is really good - I like that we have a new data set! Feel free to > merge whenever - you did not write any code that is run in the test suite, > so travis is not going to say anything useful. (I am not merging right now > because I am interested to see *if* it passes, or if travis is still > unhappy for other reasons) > > — > You are receiving this because you authored the thread. > Reply to this email directly, view it on GitHub > <#2766 (comment)>, > or mute the thread > <https://github.com/notifications/unsubscribe-auth/AA8DiC5EdwX51m8eh58lQYRfKgOwyooMks5tFRGugaJpZM4RPC7D> > . >

[AustinRochford]

If you're using format instead of f-strings, you can take the lambda out of this expression

CONVERGENCE_TITLE = lambda: "BFMI = {}\nGelman-Rubin = {}".format(bfmi, max_gr)

[AustinRochford]

Looks good overall! My only comment is that you may be able to avoid duplicating some of the model math in simulate_season with sample_ppc.

[springcoil]

I think the sample PPC could be a future PR

…

On 1 Jan 2018 15:38, "Austin Rochford" ***@***.***> wrote: Looks good overall! My only comment is that you may be able to avoid duplicating some of the model math in simulate_season with sample_ppc. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#2766 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AA8DiE2JZEGGUV6d2PjdiJf5XboomrSJks5tGO3mgaJpZM4RPC7D> .

[junpenglao]

Thanks @springcoil!

[springcoil]

Anyone know why the build is still failing?

[junpenglao]

Travis CI is not building correctly in py2.7 on master for a while now. @twiecki will contact the Travis support.