Plotly: iframes only work when pressing knit html button in rstudio

[sckott]

Or at least for some people the hook only works when pressing knit html button in rstudio.

Perhaps we need to edit the knit hook to make it work in any context.

[tdhock]

Hm, I am not an expert in writing knit hooks, so do you have any ideas? In
practice I just use the knit html button in rstudio. The other context is
just using knit2html on the R command line?

Another thing that would be nice but I don't have any idea how to do is to
change the knit hook to automatically detect and embed plotlys (rather than
having to specify plotly=TRUE), any idea how to do that?

On Fri, Apr 18, 2014 at 2:13 AM, Scott Chamberlain <[email protected]

wrote:

Or at least for some people the hook only works when pressing knit html
button in rstudio.

Perhaps we need to edit the knit hook to make it work in any context.

—
Reply to this email directly or view it on GitHubhttps://github.com/ropensci/plotly/issues/12
.

[nacnudus]

When knitting from the command line, interactive() is true.

pub$ggplotly <- function(gg=last_plot()){
    if(!is.ggplot(gg)){
      stop("gg must be a ggplot")
    }
    pargs <- gg2list(gg)
    if(interactive()){ # we are on the command line.
      resp <- do.call(pub$plotly, pargs)
      browseURL(resp$url)
      invisible(list(data=pargs, response=resp))
    }else{ # we are in knitr/RStudio.
      do.call(pub$iplot, pargs)
    }
  }

Not sure how to get around this. Interactivity seems to be a property of the session. Is it possible to create a new non-interactive session from within the console?

[ouzor]

I can not get the embedded iframes to work, tried both RStudio's 'knit html' button and knit(). Trying to replicate this: http://ropensci.org/blog/2014/04/17/plotly/.
Here's my sessionInfo()
R version 3.1.1 (2014-07-10)
Platform: x86_64-apple-darwin13.1.0 (64-bit)
locale: [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
attached base packages:
[1] stats graphics grDevices utils datasets methods base
loaded via a namespace (and not attached):
[1] digest_0.6.4 htmltools_0.2.6 rmarkdown_0.3.10 tools_3.1.1 yaml_2.1.13

[mkcor]

Hello @ouzor !

Follow the syntax below when writing your .Rmd file:

You want to set the session parameter to "knitr" in the ggplotly() call and also add the plotly=TRUE option in the code chunk where you embed a Plotly plot.

Hope this helps!

[mkcor]

@ouzor Note that after clicking "Knit HTML" I also had to "Open in Browser" in order to see the embedded Plotly plot. The plot was generated (you can check online in your Plotly account), it's just a rendering thing.

[ouzor]

Thanks @mkcor, this works! I was not able to find much documentation for the ggplotly command, is that available somewhere?

[mkcor]

@ouzor We are indeed behind with respect to documentation.
Now this is documented here: http://blog.plot.ly/post/106630252117/plot-with-ggplot2-and-plotly-within-knitr-reports
Thank you for your patience!

[ouzor]

Thanks!
I made a short comparison of different interactive visualization packages in R, including plotly, see here: http://ouzor.github.io/blog/2014/11/21/interactive-visualizations.html
I like very much the idea of turning an existing ggplot2 object to a nice interactive version. However, as a data scientist I work a lot with confidential data and can not make any of that public, which to me seems that I can not use the free version of plotly at all...

[mkcor]

Hi @ouzor

We definitely understand that’s a concern. Thanks for asking. A couple thoughts.

You control the privacy of your data, and can make your work public or keep it private (like Google Docs or GitHub).

You own your data; we don’t make an IP claim to it. We also have an on-premise version of plotly you can run on your own servers. Would any of these options work? Let me know if we can help with anything.

[ouzor]

Thanks a lot @mkcor, I'll take a closer look at these options!

[makis23]

why does this not work??
library(shiny)
library(plotly)

Rely on the 'WorldPhones' dataset in the datasets

package (which generally comes preloaded).

library(datasets)

Define a server for the Shiny app

shinyServer(function(input, output) {

# Fill in the spot we created for a plot
output$phonePlot <- renderPlotly({

# Render a barplot

gg <- barplot(WorldPhones[,input$region]*1000,
main=input$region,
ylab="Number of Telephones",
xlab="Year")
})

p <- ggplotly(gg,session="knitr")
p
})

[aniruhil]

Despite following @mkcor's fix that seemed to work for @ouzor, I run into the listed error msg when running Knit HTML and when executing the code line-by-line from inside RStudio.

``` r, plotly=TRUE
require(ggplot2)
gg <- ggplot(cars) + geom_point(aes(speed, dist))
gg

require(plotly)
py <- plotly("aniruhil", "maskedkey")
py$ggplotly(gg, session="knitr")
```

Quitting from lines 27-30 (testingplotly.Rmd)
Error in eval(expr, envir, enclos) : attempt to apply non-function
Calls: ... handle -> withCallingHandlers -> withVisible -> eval -> eval
Execution halted

[cpsievert]

The interface has changed quite a bit since this thread first started. You should only need this:

library(plotly)
gg <- ggplot(cars) + geom_point(aes(speed, dist))
ggplotly(gg)

[aniruhil]

thanks Carson