Sliders / interactions for rbokeh [R]

Hello all,

First post to this group. I started with bokeh examples on the web using python but my primary data analysis is done in R. I am hoping to use this tool for rapid data exploration, because it is potentially more attractive than my default approach using ggplot.

However, I noticed that there are no default options to create any of the more interesting interactions such as sliders, selects, and the like, as we have the in python+web interface. After some intense googling and documentation-hunting, I still came up short. I’m assuming that this functionality therefore doesn’t exist. I wanted to make sure these don’t fact in exist (beyond what is listed on http://hafen.github.io/rbokeh/#tools), and/or if there are other ways to mimic this functionality.

One particular use case in mind is I want to visualize model runs, with three dimensions of output variables (X, Y and color) and slider(s)/select(s)/other to quickly change multiple dimensions of model input parameters.

Thanks in advance,

Michael

Michael,

Bokeh does have this functionality. I beleive it can be implemented in the style of Shiny via the Bokeh Server, or setup for pure client side via the use of the Widgets without the Bokeh Server. You can find information about the Widget system in the documentation at:

http://bokeh.pydata.org/en/0.11.1/docs/user_guide/interaction.html

The layouts of the Widgets on the page is briefly discussed at:

http://bokeh.pydata.org/en/0.11.1/docs/user_guide/layout.html

The layout engine is not as friendly as Shiny yet, but it is a focus area for the developers right now. Hope this helps.

Justace

···

On Apr 2, 2016, at 12:04 AM, Michael Davidson [email protected] wrote:

Hello all,

First post to this group. I started with bokeh examples on the web using python but my primary data analysis is done in R. I am hoping to use this tool for rapid data exploration, because it is potentially more attractive than my default approach using ggplot.

However, I noticed that there are no default options to create any of the more interesting interactions such as sliders, selects, and the like, as we have the in python+web interface. After some intense googling and documentation-hunting, I still came up short. I’m assuming that this functionality therefore doesn’t exist. I wanted to make sure these don’t fact in exist (beyond what is listed on http://hafen.github.io/rbokeh/#tools), and/or if there are other ways to mimic this functionality.

One particular use case in mind is I want to visualize model runs, with three dimensions of output variables (X, Y and color) and slider(s)/select(s)/other to quickly change multiple dimensions of model input parameters.

Thanks in advance,

Michael

You received this message because you are subscribed to the Google Groups “Bokeh Discussion - Public” group.

To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].

To post to this group, send email to [email protected].

To view this discussion on the web visit https://groups.google.com/a/continuum.io/d/msgid/bokeh/281caf68-57c2-46c2-a098-45a7d4d92e25%40continuum.io.

For more options, visit https://groups.google.com/a/continuum.io/d/optout.

Hi Michael,

The interface to BokehJS is a declarative JSON specification, which means that it would be possible create and use Bokeh widgets with CustomJS callbacks from R, if the R interface supports creating the right JSON to drive BokehJS. I don't think that it does currently (for widgets), but I may be wrong. Another possibility would to have an R interface that integrates Bokeh plots well with other existing "R native" widgets. Something along these lines sounds familiar to me, but I am not the best resource. RBokeh is primarily maintained by Ryan Hafen, and I am not sure how much he is able to follow this list. I'd suggest opening an issue to discuss this on the RBokeh issue tracker:

  https://github.com/bokeh/rbokeh/issues

It is also possible (in principle) that the Bokeh Server could be extended to support creating Bokeh server applications from R scripts, i.e. that it would be possible to do:

  bokeh serve myapp.r

This is a more ambitious goal. Just having the server call out to R to get the JSON for a standalone Bokeh app seems like it would be fairly straightforward to do. The complicated issue would be to make it possible for widget callbacks to execute R code to update the app state. I'd love to see this capability, but I couldn't begin to speculate on a time frame.

Regards,

Bryan

···

On Apr 1, 2016, at 11:04 PM, Michael Davidson <[email protected]> wrote:

Hello all,

First post to this group. I started with bokeh examples on the web using python but my primary data analysis is done in R. I am hoping to use this tool for rapid data exploration, because it is potentially more attractive than my default approach using ggplot.

However, I noticed that there are no default options to create any of the more interesting interactions such as sliders, selects, and the like, as we have the in python+web interface. After some intense googling and documentation-hunting, I still came up short. I'm assuming that this functionality therefore doesn't exist. I wanted to make sure these don't fact in exist (beyond what is listed on http://hafen.github.io/rbokeh/#tools), and/or if there are other ways to mimic this functionality.

One particular use case in mind is I want to visualize model runs, with three dimensions of output variables (X, Y and color) and slider(s)/select(s)/other to quickly change multiple dimensions of model input parameters.

Thanks in advance,

Michael

--
You received this message because you are subscribed to the Google Groups "Bokeh Discussion - Public" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
To post to this group, send email to [email protected].
To view this discussion on the web visit https://groups.google.com/a/continuum.io/d/msgid/bokeh/281caf68-57c2-46c2-a098-45a7d4d92e25%40continuum.io.
For more options, visit https://groups.google.com/a/continuum.io/d/optout.