Expand pp checking

[jgabry]

First version of pp checking will be included in v1.1.0 to be released soon.

Opening this issue to keep improving and expanding pp checking capabilities on the to-do list.

[bgoodri]

Also, for a stanreg object, shinystan should just call posterior_predict rather than asking the user to select the name of the posterior predictions.

[jgabry]

Yes, good call. I will put this in the next release. Thoughts on particular pp checking plots for rstanarm models?

If I really wanted to put some time into it I could have shinystan automatically adjust various settings and plots for the different rstanarm models using the info in the stanreg object to classify the model.

On Thursday, August 13, 2015, bgoodri [email protected] wrote:

Also, for a stanreg object, shinystan should just call posterior_predict rather than asking the user to select the name of the posterior predictions.

— Reply to this email directly or view it on GitHub https://github.com/stan-dev/shinystan/issues/67#issuecomment-130890463.

[bgoodri]

I don't really know; start with just making it possible to do something with a stanreg object because it is not going to have a drop-down entry for y_rep. Rather than investing a lot of time customizing for stanreg objects, it seems that you could do a fair amount of reasonable customization just based on the class of the dependent variable.

On Thu, Aug 13, 2015 at 8:23 PM, Jonah Gabry [email protected] wrote:

Yes, good call. I will put this in the next release. Thoughts on particular pp checking plots for rstanarm models?

If I really wanted to put some time into it I could have shinystan automatically adjust various settings and plots for the different rstanarm models using the info in the stanreg object to classify the model.

On Thursday, August 13, 2015, bgoodri [email protected] wrote:

Also, for a stanreg object, shinystan should just call posterior_predict rather than asking the user to select the name of the posterior predictions.

— Reply to this email directly or view it on GitHub <https://github.com/stan-dev/shinystan/issues/67#issuecomment-130890463 .

— Reply to this email directly or view it on GitHub https://github.com/stan-dev/shinystan/issues/67#issuecomment-130900534.

[jgabry]

Yeah, definitely. Recognizing the dependent variable type and selecting the content accordingly is one of the things I've been wanting to do.

On Thu, Aug 13, 2015 at 8:27 PM, bgoodri [email protected] wrote:

I don't really know; start with just making it possible to do something with a stanreg object because it is not going to have a drop-down entry for y_rep. Rather than investing a lot of time customizing for stanreg objects, it seems that you could do a fair amount of reasonable customization just based on the class of the dependent variable.

On Thu, Aug 13, 2015 at 8:23 PM, Jonah Gabry [email protected] wrote:

Yes, good call. I will put this in the next release. Thoughts on particular pp checking plots for rstanarm models?

If I really wanted to put some time into it I could have shinystan automatically adjust various settings and plots for the different rstanarm models using the info in the stanreg object to classify the model.

On Thursday, August 13, 2015, bgoodri [email protected] wrote:

Also, for a stanreg object, shinystan should just call posterior_predict rather than asking the user to select the name of the posterior predictions.

— Reply to this email directly or view it on GitHub < https://github.com/stan-dev/shinystan/issues/67#issuecomment-130890463 .

— Reply to this email directly or view it on GitHub <https://github.com/stan-dev/shinystan/issues/67#issuecomment-130900534 .

— Reply to this email directly or view it on GitHub https://github.com/stan-dev/shinystan/issues/67#issuecomment-130901343.

[bgoodri]

At a minimum, a stanreg object can store its y in a canonical form for each type of model so it is easy to use if shinystan knows those forms.

On Thu, Aug 13, 2015 at 8:37 PM, Jonah Gabry [email protected] wrote:

Yeah, definitely. Recognizing the dependent variable type and selecting the content accordingly is one of the things I've been wanting to do.

On Thu, Aug 13, 2015 at 8:27 PM, bgoodri [email protected] wrote:

I don't really know; start with just making it possible to do something with a stanreg object because it is not going to have a drop-down entry for y_rep. Rather than investing a lot of time customizing for stanreg objects, it seems that you could do a fair amount of reasonable customization just based on the class of the dependent variable.

On Thu, Aug 13, 2015 at 8:23 PM, Jonah Gabry [email protected] wrote:

Yes, good call. I will put this in the next release. Thoughts on particular pp checking plots for rstanarm models?

If I really wanted to put some time into it I could have shinystan automatically adjust various settings and plots for the different rstanarm models using the info in the stanreg object to classify the model.

On Thursday, August 13, 2015, bgoodri [email protected] wrote:

Also, for a stanreg object, shinystan should just call posterior_predict rather than asking the user to select the name of the posterior predictions.

— Reply to this email directly or view it on GitHub < https://github.com/stan-dev/shinystan/issues/67#issuecomment-130890463 .

— Reply to this email directly or view it on GitHub < https://github.com/stan-dev/shinystan/issues/67#issuecomment-130900534 .

— Reply to this email directly or view it on GitHub <https://github.com/stan-dev/shinystan/issues/67#issuecomment-130901343 .

— Reply to this email directly or view it on GitHub https://github.com/stan-dev/shinystan/issues/67#issuecomment-130902415.