pcFactorStan

I’m posting to contribute a new case study, https://cran.r-project.org/web/packages/pcFactorStan/

Can you hook this up to your doc page? https://mc-stan.org/users/documentation/case-studies.html

Thank you.

1 Like

That was a link to a CRAN page. Was there a case study somehwere? Were you thinking about the manual, which is the only case-study-like thing I see on CRAN?

@breckbaldwin is in charge of the web site now, so it’s up to him to decide whether we want to publish manuals for software that depends on Stan as case studies. So far, we’ve concentrated on the packages that are part of our GitHub organization, stan-dev.

Yeah, the vignette

You have case studies from a similar package, edstan. I’m not even sure where to find the edstan source repo.

First off thanks for the work and packaging it all up.

The manual (https://cran.r-project.org/web/packages/pcFactorStan/vignettes/brief-manual.html) looks like a nice writeup of how to use the package.

We need to make a decision regarding package contributions to case studies. You are correct that EdStan is in there so why not?

We will be discussing this at this Thursday’s Stan meeting 11am EDT. I’ll get back after then. You are welcome to attend, link to meeting at Stan General Thursday Meeting, June 20, 2019 11 am EDT

Breck

Unfortunately I’m not available at that time. I am eager to hear any feedback.

Thank you.

@breckbaldwin What was the decision? I believe you’re still in charge of the web site.

We discussed this at the meeting and we want to create some sort of wiki or awesome * repo in the spirit of:

https://github.com/rigtorp/awesome-modern-cpp

https://github.com/rust-unofficial/awesome-rust

https://github.com/avelino/awesome-go

https://github.com/vinta/awesome-python

I think the awesome * approach works best (thanks Sean T) but now we have to find the time to build/curate it. No way I am doing it before StanCon. Any volunteers?

Breck

There are a bunch of these projects in the air:

Awesome Stan

That was the original motivation for example-models. It’s just that over time, we don’t think they’re awesome any more. I completely agree it’d be nice to have a set of reference models illustrating best practices of both statistics and computation.

Beautiful Stan

O’Reilly books has a franchise called Beautiful X, like Beautiful Testing and Beautiful Code. That’s the role our case studies play. They’re more tutorial articles than just code examples.

Validated Stan

There’s also the stat_comp_benchmarks repo that has models we believe will actually work and can be used to test algorithms. I’d like to get people off of testing with example-models and build this out instead.

Stan Contrib

The other thing we discussed was something wiki-like that we wouldn’t have to manage and people could contribute to. That’s not something we’d want to call “Beautiful Stan” or “Awesome Stan”, as we wouldn’t be vetting it.

1 Like