Online library of Stan models

[Jonah]

An issue opened by a user at the shinyStan github repo got me thinking, an experimenting, and I was able to get a slightly modified version of the shinyStan demo up and running online at

https://jgabry.shinyapps.io/shinyStan_demo/

It's cool enough to have this demo online now, but the really cool thing is that this opens the door to Stan users having an online library, so to speak, of their personal Stan models. Essentially, Stan users can deploy shinyStan apps on shinyapps.io using their own models. Each model can be a self-contained app. It takes about 2 minutes to get them online. For example, while I was typing this I uploaded a different version of the same demo model (this one using the matt trick) here

https://jgabry.shinyapps.io/shinyStan-Demo-2/

So I've just shared my Stan model with you by giving you a url!

Another thing we could do with this ability is have a library of shinyStan apps for the Stan demo models and people could view them online.

[Jonah]

By the way, it works on an iPad too.

[Daniel Lee]

Sweet!

On Friday, February 27, 2015, Jonah <jga...@gmail.com> wrote:

By the way, it works on an iPad too.

--
You received this message because you are subscribed to the Google Groups "stan development mailing list" group.
To unsubscribe from this group and stop receiving emails from it, send an email to stan-dev+u...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Bob Carpenter]

That's really awesome.

Some feature requests/questions, in no particular
order or priority:

1. The posterior summary stats is wasting a lot of
horizontal space (super premium on the web) for the
control. Is there some way to minimize that when it's
not being used? Maybe put it above or below the plot?

2. When you trim the number of entries in the table,
it'd be great to have the trim be for containers.

A nice option might be to have each parameter unfold
or collapse, as in a standard hierarchical file browser.
Then the trim might be how many to show when it's expanded.
I doubt anyone's going to have enough named parameters that
including them all is problematic.

3. For tables of numbers, there should always be right
alignment so the decimal places line up if possible. Nothing's
going to solve the parameters with different scales problems
other than scientific notation throughout.

4. The show/hide customization thing on the parameters plot
is hard to see. I think we'll need conventions as to where
these general controls go. For something this general, not
having it on the display itself makes sense to me if that's
possible.

5. I didn't understand the "ALL theta_raw" vs. "theta_raw"
option, even after poking at it a bit. "theta_raw" seems to
be a synonym for "theta[1]"?

6. The greyed out names under the table don't seem to
work as controls, even thought they look like they'll be
editable because they pop up cursors. I don't see the point
of having those. I changed the 50% to "44%" and got "no records
found". Then after I did that, nothing works at all. So maybe
a "reset table" option would be helpful. I went into view
table options and changing decimal places reset everything --- this
is probably not what you want to happen in general. Destructive
ops with no "back" button need strong warnings to users, maybe
in the form of a modal dialog box.

7. I don't know what First/Previous/Next/Last are for.
They look like live buttons but can't be clicked.

8. In the search box, what am I supposed to be searching for?
It looks like I can enter parameters? If I select "mu_raw"
and get just that, then edit the box on the bottom, it needs
to be one of the ones I'm viewing, which seems confusing.

9. What does the "Estimate" control at the top do? It doesn't
seem to do anything when I click it, yet it looks like something
clickable. The diagnose options are really sweet. I like that
you can customize the warning levels. The glossary's really
awesome, too.

10. Andrew has told me time and time again not to use "samples"
when I mean "draws". A sample is a collection of draws. I've
almost broken myself of the habit, but it's tough, because a whole
lot of the literature conflates "sample" and "draw". The "effective
sample size" is the effective size of the sample. A pure MC
sample has an effective sample size of the number of draws.
So the label should maybe be "effective sample size / draws" or
"n_eff / iterations" --- run it by Andrew and let him decide.

11. The summary of sample parameters is fantastic. Maybe an
option to include sd or intervals would make it even fantasticker.

12. How do you set the maximum lag for the correlation plots?
Oh, I see --- you've got customization for everything. That's
really cool. I still think the customization should be a bit more
front-and-center rather than in small text to the right. Maybe
in the same control box as the one that lets you enter parameter names.

- Bob

[Jonah Sol Gabry]

Bob, thanks for the detailed comments! They're really really helpful. I made some comments and asked you a few follow up questions inline below. 

On Fri, Feb 27, 2015 at 6:29 PM, Bob Carpenter <ca...@alias-i.com> wrote:

That's really awesome.

Some feature requests/questions, in no particular
order or priority:

1. The posterior summary stats is wasting a lot of
horizontal space (super premium on the web) for the
control.  Is there some way to minimize that when it's
not being used?  Maybe put it above or below the plot?

Yup, good idea. I'll try out a few things and see what works well. 

 

2.  When you trim the number of entries in the table,
it'd be great to have the trim be for containers.

A nice option might be to have each parameter unfold
or collapse, as in a standard hierarchical file browser.
Then the trim might be how many to show when it's expanded.
I doubt anyone's going to have enough named parameters that
including them all is problematic.

That would be cool. I'm still familiarizing myself with the DataTables library, so I'm not entirely sure how to do that, but I'll look into it. 

 

3.  For tables of numbers, there should always be right
alignment so the decimal places line up if possible.  Nothing's
going to solve the parameters with different scales problems
other than scientific notation throughout.

Very true. I'll fix the alignment. 

 

4. The show/hide customization thing on the parameters plot
is hard to see.  I think we'll need conventions as to where
these general controls go.  For something this general, not
having it on the display itself makes sense to me if that's
possible.

Yeah, I was trying to figure out where to put it and then I realized I was spending too much time fiddling with it so I just stuck it up there at the top right. But I agree that it's not great. What about sticking it up in the well panel to the right of the CI slider?  

 

5.  I didn't understand the "ALL theta_raw" vs. "theta_raw"
option, even after poking at it a bit.  "theta_raw" seems to
be a synonym for "theta[1]"?

The way it should work is that if x is a vector (or matrix, etc) then "ALL x" will select all the elements so you don't have to add each of them individually. So "ALL theta_raw" should select all 8 of the theta_raw parameters. Is that not what's happening when you use it? 

 

6.  The greyed out names under the table don't seem to
work as controls, even thought they look like they'll be
editable because they pop up cursors.  I don't see the point
of having those.  I changed the 50% to "44%" and got "no records
found".  Then after I did that, nothing works at all.  So maybe
a "reset table" option would be helpful. 

Those controls show up automatically and I forgot to try to get rid of them. I just tried using them and, while they're not particularly useful, I haven't had the problem where everything stops working. Either way, I think I should just try to remove them. 

 

I went into view
table options and changing decimal places reset everything --- this
is probably not what you want to happen in general.  Destructive
ops with no "back" button need strong warnings to users, maybe
in the form of a modal dialog box.

Good catch. I completely missed that and I definitely don't want it to reset everything when you change the decimals. I assume I'll be able to make it so that it doesn't reset, but if not I'll add a warning.  

 

7.  I don't know what First/Previous/Next/Last are for.
They look like live buttons but can't be clicked.

They should definitely be clickable in some cases. For example, if it's set to show 10 entries per page and you have 50 entries in the table you should be able to use those buttons to move between pages. But you won't be able to click "First" if you're on the first page or "Last" if you're on the last, etc. And you won't be able to click any of them if you're displaying all the entries on one page. Maybe there's a way to only have them show up under "clickable circumstances", so to speak. 

 

8.  In the search box, what am I supposed to be searching for?
It looks like I can enter parameters?  If I select "mu_raw"
and get just that, then edit the box on the bottom, it needs
to be one of the ones I'm viewing, which seems confusing.

Yeah the search box is for parameter names, and you can use regular expressions. It's not always useful, but, for example if you have a lot of betas and a lot of sigma_betas then it's nice to be able to type ^beta and get only the betas without the sigma_betas. If I get rid of those boxes on the bottom will it be less confusing?

 

9.  What does the "Estimate" control at the top do?  It doesn't
seem to do anything when I click it, yet it looks like something
clickable. 

Oh, I can see why that's confusing. It's just the name for the tab that has the parameters plot and summary stats table (as opposed to the diagnostics, etc). I kind of like "Diagnose" and "Explore", so I went with "Estimate" the verb. But I could change it to "Estimates" or something else if anyone has a better idea.  

 

10.  Andrew has told me time and time again not to use "samples"
when I mean "draws".  A sample is a collection of draws.  I've
almost broken myself of the habit, but it's tough, because a whole
lot of the literature conflates "sample" and "draw".  The "effective
sample size" is the effective size of the sample.  A pure MC
sample has an effective sample size of the number of draws.
So the label should maybe be "effective sample size / draws" or
"n_eff / iterations" --- run it by Andrew and let him decide.

Thanks for pointing this out. I definitely tend to use "samples" and "draws" (and sometimes "simulations") interchangeably. 

 

11.  The summary of sample parameters is fantastic.  Maybe an
option to include sd or intervals would make it even fantasticker.

Well I definitely won't settle for fantastic when fantasticker is an option. Adding sds and/or intervals is definitely doable. 

 

12.  How do you set the maximum lag for the correlation plots?
Oh, I see --- you've got customization for everything.  That's
really cool.  I still think the customization should be a bit more
front-and-center rather than in small text to the right.  Maybe
in the same control box as the one that lets you enter parameter names.

That's how I used to have things, but for the plots with a lot of customization options (like the main parameters plot) the screen would get very cluttered. That's why I went with the separate panels that get transparent when you're not hovering over them. But maybe that's not the right solution. 

Thanks again for all the comments!

Jonah

P.S. Out of curiosity which web browser were you using? And were you using a laptop or trying it out on an iPad?

[Bob Carpenter]

> On Feb 28, 2015, at 12:36 PM, Jonah Sol Gabry <jga...@gmail.com> wrote:
> ...

> 4. The show/hide customization thing on the parameters plot
> is hard to see. I think we'll need conventions as to where
> these general controls go. For something this general, not
> having it on the display itself makes sense to me if that's
> possible.
>
> Yeah, I was trying to figure out where to put it and then I realized I was spending too much time fiddling with it so I just stuck it up there at the top right. But I agree that it's not great. What about sticking it up in the well panel to the right of the CI slider?

That sounds good, though I might put it up front so
it's in the same physical space on every page. Maybe
a small-ish button? I don't know what kind of GUIs you
use.

By the way, I would highly recommend two books that changed
the way I thought about GUI design (though I've read a lot of
them):

* Jeff Johnson. GUI Bloopers.

I read the first edition in book form and the second in
draft form (the editor contacted me based on my enthusiastic
Amazon feedback, which was mighty surprising). I also like

* Stephen Krug. Don't Make Me Think.

I think I read the first edition and it's now up to the 3rd. And
also this one about general design:

* Robin Williams. Non-Designers' Design Book. (not web specific,
but great advice in general on layout)

It's also many editions past where I first read it. I'm also a huge
fan of

* Hunt and Thomas. The Pragmatic Programmer.

for general programming and development advice (though you may be
well beyond needing to read that --- we've implemented most of
the recommended practices in Stan's workflow).

If you want to buy them, I'm guessing Andrew won't mind reimbursing
you (though you should ask first). I keep giving away my own copies.

> 5. I didn't understand the "ALL theta_raw" vs. "theta_raw"
> option, even after poking at it a bit. "theta_raw" seems to
> be a synonym for "theta[1]"?
>
> The way it should work is that if x is a vector (or matrix, etc) then "ALL x" will select all the elements so you don't have to add each of them individually. So "ALL theta_raw" should select all 8 of the theta_raw parameters. Is that not what's happening when you use it?

I think it's the fact that you have "ALL theta_raw" and "theta_raw"
as options --- it's really the second one that's confusing because
it seems to be a synonym for "theta_raw[1]".

> 6. The greyed out names under the table don't seem to
> work as controls, even thought they look like they'll be
> editable because they pop up cursors. I don't see the point
> of having those. I changed the 50% to "44%" and got "no records
> found". Then after I did that, nothing works at all. So maybe
> a "reset table" option would be helpful.
>
> Those controls show up automatically and I forgot to try to get rid of them. I just tried using them and, while they're not particularly useful, I haven't had the problem where everything stops working. Either way, I think I should just try to remove them.

I agree.

> 7. I don't know what First/Previous/Next/Last are for.
> They look like live buttons but can't be clicked.
>
> They should definitely be clickable in some cases. For example, if it's set to show 10 entries per page and you have 50 entries in the table you should be able to use those buttons to move between pages. But you won't be able to click "First" if you're on the first page or "Last" if you're on the last, etc. And you won't be able to click any of them if you're displaying all the entries on one page. Maybe there's a way to only have them show up under "clickable circumstances", so to speak.

Or have them be clearly greyed out, which usually requires
contrast with non-greyed-out elements.

> 8. In the search box, what am I supposed to be searching for?
> It looks like I can enter parameters? If I select "mu_raw"
> and get just that, then edit the box on the bottom, it needs
> to be one of the ones I'm viewing, which seems confusing.
>
> Yeah the search box is for parameter names, and you can use regular expressions. It's not always useful, but, for example if you have a lot of betas and a lot of sigma_betas then it's nice to be able to type ^beta and get only the betas without the sigma_betas. If I get rid of those boxes on the bottom will it be less confusing?

Yes. And maybe a label for "search parameters"? But that's
long enough you'd need to align it on top of rather than in
front of the search box. I'm not a huge fan of having text
inside the box that goes away when you add your own.

> 9. What does the "Estimate" control at the top do? It doesn't
> seem to do anything when I click it, yet it looks like something
> clickable.
>
> Oh, I can see why that's confusing. It's just the name for the tab that has the parameters plot and summary stats table (as opposed to the diagnostics, etc). I kind of like "Diagnose" and "Explore", so I went with "Estimate" the verb. But I could change it to "Estimates" or something else if anyone has a better idea.

I actually figured that out when I went to other tabs and then
went back. I think it's OK as is.

> 11. The summary of sample parameters is fantastic. Maybe an
> option to include sd or intervals would make it even fantasticker.
>
> Well I definitely won't settle for fantastic when fantasticker is an option. Adding sds and/or intervals is definitely doable.

The benefits of linguistics training.

> 12. How do you set the maximum lag for the correlation plots?
> Oh, I see --- you've got customization for everything. That's
> really cool. I still think the customization should be a bit more
> front-and-center rather than in small text to the right. Maybe
> in the same control box as the one that lets you enter parameter names.
>
> That's how I used to have things, but for the plots with a lot of customization options (like the main parameters plot) the screen would get very cluttered. That's why I went with the separate panels that get transparent when you're not hovering over them. But maybe that's not the right solution.

I'm not sure --- I would prefer something that was always in
the same place and would pop up controls. Hovering partial
transparency usually confuses me (like in the default Mac terminal).

> Thanks again for all the comment

> Jonah
>
> P.S. Out of curiosity which web browser were you using? And were you using a laptop or trying it out on an iPad?

I was using Safari on my Mac. Everything worked just fine.

I also have an iPad if you'd like me to try that. We can probably
get you one of those, too, if you don't have one and are going to
be developing visualizations for it. I could also try it on my iPhone.
For that amount of money, we need to ask Andrew because it's coming out of
his slush funds (our grants don't have any free money for books, computers,
etc.)

- Bob

[Ben Goodrich]

This looks great. To me, the biggest eyesore with stanfit objects and everything that goes with them is the numbering of parameters, as in

##             mean se_mean  sd   2.5%    25%    50%    75%  97.5% n_eff Rhat
## beta[1]      1.3     0.1 3.4   -5.3   -1.0    1.3    3.6    8.0  1518    1
## beta[2]      0.1     0.0 0.8   -1.5   -0.4    0.1    0.6    1.7  2526    1
## beta[3]     -0.7     0.0 1.5   -3.8   -1.7   -0.7    0.2    2.2  1435    1
## beta[4]     -2.1     0.0 1.5   -5.0   -3.1   -2.1   -1.1    0.9  1218    1
## beta[5]     -1.7     0.0 1.4   -4.5   -2.6   -1.7   -0.9    1.1  1162    1
## beta[6]     -3.1     0.0 1.5   -6.1   -4.1   -3.1   -2.2   -0.1  1138    1
## beta[7]     -5.0     0.0 1.8   -8.5   -6.2   -5.0   -3.9   -1.5  1342    1
## beta[8]      1.1     0.0 1.0   -1.0    0.4    1.0    1.7    3.1  1610    1
## beta[9]      0.6     0.0 1.1   -1.6   -0.1    0.6    1.3    2.6  1406    1
## beta[10]     1.0     0.0 1.2   -1.5    0.2    1.0    1.8    3.3  1404    1
## beta[11]    -0.3     0.0 1.4   -3.0   -1.2   -0.3    0.7    2.5  1360    1
## beta[12]    -1.1     0.0 1.6   -4.1   -2.1   -1.1    0.0    2.0  1484    1
## beta[13]    -0.2     0.0 0.4   -0.9   -0.5   -0.2    0.0    0.5  1753    1
## sigma        1.9     0.0 0.2    1.5    1.7    1.8    2.0    2.3  1917    1
## log_prior  -13.0     0.0 0.3  -13.6  -13.1  -12.9  -12.8  -12.6  1121    1
## lp__      -126.6     0.1 3.0 -133.4 -128.4 -126.2 -124.4 -121.9  1108    1

and the user is trying to remember what beta[7] corresponds to. With the rstan on GitHub, it is technically possible to reassign names now, but you have to know what you are doing and it is difficult to then only print a subset of the renamed parameters

names(post) <- c(paste0("z_", colnames(X)), "sigma", colnames(X), "log_prior", as.character(PErisk$country), "lp__")
> post
Inference for Stan model: template.
4 chains, each with iter=2000; warmup=1000; thin=1;
post-warmup draws per chain=1000, total post-warmup draws=4000.

                      mean se_mean   sd    2.5%     25%     50%     75%   97.5% n_eff Rhat
z_(Intercept)         0.13    0.01 0.35   -0.56   -0.11    0.13    0.36    0.81  1712    1
z_courts1             0.01    0.00 0.08   -0.15   -0.04    0.01    0.06    0.17  2086    1
z_prsexp21           -0.08    0.00 0.15   -0.36   -0.17   -0.08    0.02    0.21  1688    1
z_prsexp22           -0.21    0.00 0.14   -0.51   -0.31   -0.22   -0.12    0.08  1558    1
z_prsexp23           -0.18    0.00 0.13   -0.44   -0.27   -0.18   -0.09    0.08  1422    1
z_prsexp24           -0.32    0.00 0.14   -0.60   -0.42   -0.32   -0.23   -0.03  1354    1
z_prsexp25           -0.51    0.00 0.17   -0.84   -0.62   -0.51   -0.40   -0.16  1439    1
z_prscorr21           0.11    0.00 0.11   -0.10    0.04    0.11    0.17    0.32  1561    1
z_prscorr22           0.06    0.00 0.11   -0.16   -0.02    0.06    0.13    0.28  1180    1
z_prscorr23           0.09    0.00 0.12   -0.15    0.01    0.10    0.17    0.33  1160    1
z_prscorr24          -0.03    0.00 0.14   -0.30   -0.12   -0.03    0.06    0.24  1268    1
z_prscorr25          -0.11    0.00 0.15   -0.41   -0.22   -0.12   -0.01    0.19  1249    1
z_gdpw2              -0.02    0.00 0.04   -0.09   -0.05   -0.02    0.00    0.05  1781    1
sigma                 1.86    0.00 0.19    1.53    1.72    1.85    1.97    2.27  2066    1
(Intercept)           1.29    0.08 3.50   -5.65   -1.10    1.31    3.62    8.07  1712    1
courts1               0.11    0.02 0.82   -1.52   -0.45    0.12    0.65    1.70  2086    1
prsexp21             -0.77    0.04 1.46   -3.64   -1.74   -0.79    0.19    2.07  1688    1
prsexp22             -2.13    0.04 1.44   -5.06   -3.07   -2.16   -1.18    0.76  1558    1
prsexp23             -1.80    0.04 1.34   -4.45   -2.70   -1.77   -0.92    0.82  1422    1
prsexp24             -3.19    0.04 1.44   -6.00   -4.16   -3.18   -2.25   -0.32  1354    1
prsexp25             -5.06    0.05 1.71   -8.43   -6.16   -5.09   -3.96   -1.62  1439    1
prscorr21             1.07    0.03 1.06   -0.97    0.36    1.05    1.75    3.19  1561    1
prscorr22             0.58    0.03 1.10   -1.57   -0.16    0.57    1.29    2.83  1180    1
prscorr23             0.95    0.04 1.22   -1.50    0.13    0.95    1.75    3.34  1160    1
prscorr24            -0.31    0.04 1.39   -3.00   -1.25   -0.30    0.60    2.42  1268    1
prscorr25            -1.13    0.04 1.55   -4.12   -2.18   -1.15   -0.06    1.94  1249    1
gdpw2                -0.20    0.01 0.37   -0.92   -0.46   -0.21    0.04    0.53  1781    1
log_prior           -12.97    0.01 0.27  -13.65  -13.10  -12.92  -12.78  -12.60  1180    1
Argentina            -0.51    0.03 2.13   -4.61   -1.91   -0.52    0.92    3.59  3890    1
Australia            -6.04    0.03 2.04  -10.00   -7.42   -6.06   -4.69   -1.98  4000    1
Austria              -5.98    0.03 2.02   -9.98   -7.33   -5.97   -4.64   -2.04  3948    1
Bangladesh           -1.22    0.04 2.13   -5.34   -2.65   -1.20    0.23    2.92  3605    1
Belgium              -6.09    0.03 2.04  -10.06   -7.43   -6.11   -4.74   -2.17  3876    1
Bolivia              -0.49    0.05 2.35   -5.12   -2.00   -0.50    1.08    4.08  2441    1
Botswana             -2.62    0.03 1.98   -6.53   -3.90   -2.64   -1.29    1.22  4000    1
Brazil               -2.76    0.03 1.94   -6.54   -4.11   -2.74   -1.47    1.09  4000    1
Burma                -0.90    0.03 2.02   -4.73   -2.28   -0.91    0.47    3.13  3880    1
Cameroon             -1.11    0.03 2.00   -5.08   -2.43   -1.11    0.24    2.70  4000    1
Canada               -6.89    0.03 1.95  -10.68   -8.21   -6.91   -5.62   -2.97  4000    1
Chile                -1.72    0.03 2.02   -5.75   -3.07   -1.70   -0.38    2.34  3909    1
Colombia             -1.83    0.03 2.00   -5.74   -3.14   -1.82   -0.49    2.12  3898    1
Congo-Kinshasa       -0.97    0.04 2.18   -5.29   -2.43   -0.94    0.48    3.16  3608    1
Costa Rica           -2.61    0.04 2.14   -6.78   -4.07   -2.57   -1.24    1.71  3603    1
Cote d'Ivoire        -2.89    0.03 2.04   -6.92   -4.24   -2.93   -1.58    1.07  3781    1
Denmark              -6.90    0.03 1.97  -10.70   -8.20   -6.89   -5.63   -3.03  4000    1
Dominican Republic   -2.07    0.03 1.98   -6.07   -3.36   -2.06   -0.74    1.76  4000    1
Ecuador              -1.69    0.03 2.05   -5.70   -3.03   -1.72   -0.37    2.46  4000    1
Finland              -6.83    0.03 1.98  -10.71   -8.17   -6.81   -5.52   -3.01  3988    1
Gambia, The          -2.91    0.03 2.06   -6.91   -4.29   -2.86   -1.56    1.04  4000    1
Ghana                -1.32    0.03 2.09   -5.39   -2.72   -1.35    0.04    2.82  3713    1
Greece               -1.54    0.04 2.08   -5.71   -2.90   -1.52   -0.14    2.57  3429    1
Hungary              -2.77    0.03 2.00   -6.84   -4.08   -2.75   -1.45    1.05  4000    1
India                -2.98    0.03 2.12   -7.20   -4.36   -2.98   -1.49    1.08  4000    1
Indonesia            -2.27    0.04 2.12   -6.39   -3.68   -2.27   -0.84    1.98  3031    1
Iran                 -0.01    0.05 2.29   -4.44   -1.54   -0.04    1.52    4.54  2577    1
Ireland              -5.99    0.03 2.00   -9.88   -7.35   -6.01   -4.65   -2.06  4000    1
Israel               -4.27    0.03 2.16   -8.49   -5.73   -4.25   -2.84   -0.04  4000    1
Italy                -3.00    0.03 2.03   -7.06   -4.36   -2.96   -1.64    1.02  3839    1
Japan                -5.98    0.03 1.99   -9.90   -7.34   -5.97   -4.62   -2.12  4000    1
Kenya                -1.84    0.03 2.06   -5.92   -3.21   -1.78   -0.48    2.24  4000    1
Korea, South         -2.75    0.04 2.15   -6.98   -4.20   -2.75   -1.29    1.40  3495    1
Malawi               -1.00    0.04 2.16   -5.17   -2.38   -1.02    0.43    3.30  3148    1
Malaysia             -2.73    0.03 1.98   -6.62   -4.03   -2.73   -1.42    1.04  4000    1
Mexico               -2.26    0.03 2.04   -6.35   -3.55   -2.26   -0.92    1.72  4000    1
Morocco              -1.25    0.03 1.98   -5.07   -2.57   -1.26    0.09    2.76  4000    1
New Zealand          -6.89    0.03 1.96  -10.72   -8.21   -6.92   -5.57   -3.05  4000    1
Nigeria               0.04    0.03 2.05   -3.96   -1.32    0.04    1.40    4.11  3746    1
Norway               -6.89    0.03 2.01  -10.82   -8.22   -6.86   -5.53   -2.98  3412    1
Papua New Guinea     -2.90    0.03 2.05   -6.81   -4.29   -2.90   -1.51    1.19  4000    1
Paraguay             -2.23    0.04 2.12   -6.48   -3.68   -2.21   -0.83    1.96  2756    1
Philippines          -0.16    0.03 2.04   -4.09   -1.58   -0.19    1.24    3.83  4000    1
Poland               -1.33    0.03 2.04   -5.38   -2.69   -1.30    0.04    2.61  4000    1
Portugal             -2.78    0.03 2.00   -6.74   -4.11   -2.77   -1.45    1.16  3909    1
Sierra Leone         -1.00    0.03 1.98   -4.86   -2.30   -1.01    0.29    2.94  3623    1
Singapore            -6.84    0.03 1.97  -10.64   -8.18   -6.88   -5.54   -2.86  4000    1
South Africa         -2.73    0.03 2.11   -6.93   -4.16   -2.74   -1.25    1.33  4000    1
Spain                -4.80    0.03 2.11   -8.78   -6.18   -4.82   -3.39   -0.60  4000    1
Sri Lanka            -2.01    0.03 1.98   -5.85   -3.36   -2.01   -0.68    1.81  4000    1
Sweden               -4.98    0.03 2.15   -9.27   -6.39   -4.96   -3.53   -0.83  4000    1
Switzerland          -6.88    0.03 1.97  -10.75   -8.23   -6.88   -5.58   -2.92  4000    1
Syria                -0.39    0.03 2.13   -4.77   -1.79   -0.39    1.04    3.80  4000    1
Thailand             -1.68    0.03 1.97   -5.48   -2.99   -1.69   -0.39    2.27  3634    1
Togo                 -2.29    0.03 2.06   -6.30   -3.68   -2.29   -0.88    1.73  4000    1
Tunisia              -2.07    0.03 2.00   -6.09   -3.40   -2.04   -0.74    1.88  3943    1
Turkey               -1.75    0.03 2.01   -5.71   -3.10   -1.75   -0.41    2.12  3799    1
United Kingdom       -6.87    0.03 2.00  -10.87   -8.18   -6.88   -5.55   -2.92  3736    1
Uruguay              -2.20    0.03 2.00   -6.14   -3.53   -2.19   -0.85    1.67  4000    1
Venezuela            -1.80    0.03 2.05   -5.76   -3.19   -1.80   -0.41    2.28  3766    1
Zambia               -1.07    0.03 1.94   -4.98   -2.43   -1.04    0.26    2.64  4000    1
Zimbabwe             -1.59    0.03 2.01   -5.64   -2.95   -1.57   -0.20    2.29  4000    1
lp__               -126.51    0.09 2.91 -133.23 -128.20 -126.14 -124.41 -121.81  1099    1

so if shinyStan (whether offline or online) had a more convenient way to do this, I think it would help a lot.

Ben

[Jonah Sol Gabry]

I've started creating issues on GitHub for your suggestions. GitHub issue tracking really is a great feature.

Thanks for the book suggestions too! And good timing. I was just thinking the other day that I could use some new references on these topics. The only one I've heard of is The Pragmatic Programmer, but I haven't read it (only heard good things though). The others I'm not familiar with. I look forward to checking them out. 

--
You received this message because you are subscribed to a topic in the Google Groups "stan development mailing list" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/stan-dev/ISiJ-re5MFQ/unsubscribe.
To unsubscribe from this group and all its topics, send an email to stan-dev+u...@googlegroups.com.

[Jonah Sol Gabry]

Thanks Ben. I agree, this would be a great feature. I'll open an issue on GitHub and bump this up toward the top of my to-do list. I think it would be really helpful to have this functionality. 

[Ben Goodrich]

The other killer feature for the shinyapps version would be to have a button where people could upload the .csv files produced by whatever interface to Stan they happened to use. Then call read_stan_csv to create a stanfit object and launch_shinystan() on that.

Ben

[Bob Carpenter]

If you did that, it wouldn't just be restricted to Stan.

- Bob

> On Mar 1, 2015, at 10:35 PM, Ben Goodrich <goodri...@gmail.com> wrote:
>
> The other killer feature for the shinyapps version would be to have a button where people could upload the .csv files produced by whatever interface to Stan they happened to use. Then call read_stan_csv to create a stanfit object and launch_shinystan() on that.
>
> Ben
>

> --

> You received this message because you are subscribed to the Google Groups "stan development mailing list" group.

> To unsubscribe from this group and stop receiving emails from it, send an email to stan-dev+u...@googlegroups.com.

[Bob Carpenter]

And by the way, GUI Bloopers isn't just about bloopers.
It just uses them for contrast.

- Bob

> You received this message because you are subscribed to the Google Groups "stan development mailing list" group.

> To unsubscribe from this group and stop receiving emails from it, send an email to stan-dev+u...@googlegroups.com.

[Jonah Sol Gabry]

On Sun, Mar 1, 2015 at 6:35 AM, Ben Goodrich <goodri...@gmail.com> wrote:

The other killer feature for the shinyapps version would be to have a button where people could upload the .csv files produced by whatever interface to Stan they happened to use. Then call read_stan_csv to create a stanfit object and launch_shinystan() on that.

Absolutely! And unless there's something I'm not thinking of, it shouldn't be hard to implement.