allow for additional plotting options

[wds15]

It would be great to let users customize more the plots from bayesplot to make them ready for presentations or other non-screen usage.

A functionality like rstan_gg_options would be great to have which allows for specification of options which are beyond ggplot theme options and go directly to the used geoms.

[jgabry]

I agree that would be good to do. I might not get to it before the initial
CRAN release but definitely on the short term to-do list.

On Thursday, November 3, 2016, wds15 notifications@github.com wrote:

It would be great to let users customize more the plots from bayesplot to
make them ready for presentations or other non-screen usage.

A functionality like rstan_gg_options would be great to have which allows
for specification of options which are beyond ggplot theme options and go
directly to the used geoms.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
#45, or mute the thread
https://github.com/notifications/unsubscribe-auth/AHb4Q5pbQ4bbxbGJH70bIZuIFW_jLvuPks5q6eeDgaJpZM4KoYeq
.

[jgabry]

@wds15 Which particular additional options do you think would be most useful?

[wds15]

Argh... these notifications... sorry for the delay.

The main goal is to be able to adjust for using the plots in different contexts. Namely:

• data exploration on screen - most often small sizes
• presentation - things need to be large and bold
• publication - few colours, moderate size

So I guess we need point/line/font sizes for this to happen. Optimally the three use-cases I outlined are easily set by the user via an example in the help or similar.

[bob-carpenter]

The moderate size thing is the trickiest. Ideally we'd just keep PDFs, but R generates PDFs proportional in size to the number of data points for things like scatter plots or line plots. I think you'll also find that each publication has different standards, so that may not be a reasonable thing to try to anticipate in advance. I don't agree about presentations needing to be bold if you mean the font weight. But everything needs to be readable by the audience. And think about us old farts. Don't use 6pt fonts, even for exploration! - Bob

…

On Jan 3, 2017, at 4:40 AM, wds15 ***@***.***> wrote: Argh... these notifications... sorry for the delay. The main goal is to be able to adjust for using the plots in different contexts. Namely: • data exploration on screen - most often small sizes • presentation - things need to be large and bold • publication - few colours, moderate size So I guess we need point/line/font sizes for this to happen. Optimally the three use-cases I outlined are easily set by the user via an example in the help or similar. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.

[jsta]

Maybe I just can't find it but there doesn't seem to be a way to change the legend text on ppc_intervals to anything but y and y_{rep}.

[mcol]

@jsta I've opened issue #157 for this.

[avehtari]

but definitely on the short term to-do list.

Ping @jgabry

I'm using mcmc_scatter as it plots also divergences, but changing alpha doesn't work for the default shape=19, and I really would like to be able to change the shape. It is possible to change the shape of the divergence points, but not the main points.

[jgabry]

I'm using mcmc_scatter as it plots also divergences, but changing alpha doesn't work for the default shape=19, and I really would like to be able to change the shape. It is possible to change the shape of the divergence points, but not the main points.

@behramulukir maybe this is something you could add if you're interested?

[behramulukir]

Yes, I can take care of it. I think adding a set of optional parameters to scatter_style_np to control the visuals of the main points would be enough, right?

[jgabry]

Thanks @behramulukir. I think in this case you'd want to add a shape argument to mcmc_scatter itself. The scatter_style_np function is intended to just control the appearance of the NUTS sample parameters that are plotted (e.g. the divergences).

[behramulukir]

Alright, makes sense. I'll work on this