Matthew Kay comments

Results 308 comments of


                                            Matthew Kay

Pretty printing MCSE

This seems related to (or an alternative to?) #44

Coding matrix/vector variables in tidy form

Unless I'm misunderstanding that is pretty close to what spread_draws / gather_draws are for in tidybayes (amongst other things).

Coding matrix/vector variables in tidy form

Yeah! Long term I'd like to replace `tidybayes::tidy_draws()` with (or have it be a wrapper around) `posterior::as_draws_df()`. The other functions in tidybayes use `tidy_draws()` internally, so once that change happens...

Coding matrix/vector variables in tidy form

> One way of differentiating between tidybayes and posterior functions could be whether data or ploting is involved in addition to posterior draws. So whatever only handles draws could (potentially)...

Coding matrix/vector variables in tidy form

I agree the split between posterior / tidybayes is clear now. Re: the original purpose of this issue (array reshaping utility functions), I wonder if @statscats @lauken13 @jgabry have thoughts...

Coding matrix/vector variables in tidy form

Thanks for starting this! `get_indices()` is a little generic. Maybe `parse_variable_indices()`? Also, unless I've missed something the `.dimnames` information is not returned anywhere. This is needed inside as_draws_rvars to rename...

Coding matrix/vector variables in tidy form

> Note that it treats the columns independently in the sense that, despite that the first column is not one-indexed (and therefore doesn't get filled in with the full sequence...

Coding matrix/vector variables in tidy form

Awesome, thanks! Will do

Improve support (or documentation) for weighted draws

Thanks, I think there's a couple of different issues here. > weight_draws() is not defined for draws_rvars It is, I think you were using it on an `rvar` not a...

Improve support (or documentation) for weighted draws

I looked into this more for `rvar`s and in principle I think it is possible without *too* much surgery. Since we already have a common function that is used to...