Turing still thinking of “canned” packages?

[nickeubank]

I was recently watching the 2019 Turing.jl presentation from Julia con, and saw that there was discussion of creating some high-level model interfaces in the spirit of our rstanarm. I also found it mentioned in two very old issues: #460 #774.

As somebody just moving into Bayesian analysis and hoping to be able to do it within Julia, this seems like it would be a great feature. Is this still something the project is pursuing?

(Let me know if this belongs on discourse and not here)

[cpfiffer]

I had some demo work on that here, but it never made it into a full production. There's been some interest from people in a more focused development process for it, but I don't have a lot of time nowadays.

[nickeubank]

That's great! That's exactly the kind of think I think would be really helpful for beginners.

[yebai]

Please see https://github.com/TuringLang/TuringGLM.jl