xlab & ylab getting cut off using ggplotly()

[akraemer007]

Hello,

I'm new to Plotly and am very excited to use it's flash visualizations and the tooltip. I'm running into an issue with my x & y labels getting cut off when I run the viz with ggplotly -- whereas it works fine when I just run the ggplot().

Below is the code and viz. I appreciate the help in advance!

Andrew

Here's my code:

testPlot <-
ggplot(TopCarePaths.Req,aes(x = calc_appears_offset, y = total.completed_sum/total.completed_and_expired_sum ) )+
  geom_point( aes( shape = is_form,  color = PrePostFlag, text = paste("Title:", title_full) ), alpha = 2/5, size = 5) + #dimensions of CarePath
  # geom_point( aes( text = paste("Title:", title_full) ) ) + # changes the tooltip
  # geom_rug(sides="b", position='jitter') + #adds rug, which counts number of points. Added so only applied to X-axis & spread out with jitter
  guides( alpha = FALSE ) + #hides unneccssary legend for hard coded transparancy
  facet_grid( organization_parent_name~. ) + #Facets graph horizontally
  scale_x_continuous( limits = c( -50, 50 ), breaks = seq( -50, 50, 5 ) ) + #controls X axis tick marks
  scale_y_continuous(labels = scales::percent) + #quickly changes y axis to %
  geom_vline( xintercept = 0 ) + #adds solid line to chart
  geom_text( aes( x = 0, label = "Surgery\n", y = .2), family = 'sans', angle=0,  color = 'gray34', size = 3, alpha = 1/5 ) + #puts text next to line for surg date
  theme_bw() + #changes to a simpler, easy to read theme
  theme( strip.background = element_blank() ) + #removes ugly gray from name area of facet
  ggtitle( 'Content Completion by Across Full CarePath' ) +
  ylab('Completion %') +
  xlab( 'Days Appears Offset from Surgery' )
ggplotly()

[bhive01]

I'm seeing this too on my plots lately, but it appears to be related to the window size/ratio.
Here's a MRE:

require(ggplot2)
require(plotly)

randomName <- function(n) {
    random.string <- rep(NA, n)
    randomizeString <- function(x) {
        a <-sample(letters, 1, replace = TRUE)
        return(a)
    }
    return(paste(c(sapply(random.string, randomizeString, simplify = TRUE)), collapse = ""))
}

xvar = replicate(10, randomName(15))

df1 <- data.frame(x = xvar, y = xvar, z = runif(10) )

g = ggplot(df1, aes(x,y,fill=z)) +
        geom_tile() 

ggplotly(g)

Squished, the axis name overlaps with the axis labels:

nearly 1:1 everything looks ok

longer and the axis name goes off the end

[andy-ewing]

I'm having similar issues since the latest updates to R, RStudio, and plotly packages. It may have something to do with the scales package and/or faceting?

[jasdumas]

I'm experiencing the same issue with making an unstacked bar chart and the y-axis label is getting cut-off similarly to @Kidman007. I have tried some theme() adjustments within ggplot but with no success.

Has there been any solutions yet?

[timelyportfolio]

Looking at it now, but I am still not as familiar with the code as I would like to be. It appears these lines are most likely to hold the answer. I'll dig deeper.

[timelyportfolio]

When I reproduce the plot with the example from @bhive01, I get an annotation for "x" with yref equal to paper which means the annotation will be positioned with reference to the chart paper box on a scale (0,1). The negative value provided (-0.07990868) will scale linearly and inversely based on the size of the plot, which explains why when the plot is small, the x is too high, and when the plot is large, x is too low/off the screen.

I'll ping @cpsievert to see if there is an ideal solution, or if we should resort to a hack.

[cpsievert]

Yea, I'm not sure the best way to approach this problem. What I really need is a way to place annotations on a (0-1) scale of the entire plot, or on the absolute pixel scale. Any idea if that would be easy to add to plotly.js?

[timelyportfolio]

@cpsievert, do you remember why in lines this comment was made?

        # draw axis titles as annotations
        # (plotly.js axis titles aren't smart enough to dodge ticks & text)

I am guessing facets, but could be way off here.

Let's hack on the example given above by defining $x$layout$xaxis$title.

require(ggplot2)
require(plotly)

randomName <- function(n) {
  random.string <- rep(NA, n)
  randomizeString <- function(x) {
    a <-sample(letters, 1, replace = TRUE)
    return(a)
  }
  return(paste(c(sapply(random.string, randomizeString, simplify = TRUE)), collapse = ""))
}

xvar = replicate(10, randomName(15))

df1 <- data.frame(x = xvar, y = xvar, z = runif(10) )

g = ggplot(df1, aes(x,y,fill=z)) +
  geom_tile() 

gpw <- as.widget(ggplotly(g))

gpw$x$layout$xaxis$title = "plotly's x"

gpw

[jasdumas]

I've tried hacking with gpw$sizingPolicy$padding = 10 but it's still cut-off. For reference my plotly viz is on shiny where the yaxis is being cut-off by the navbarPage.

However I just downloaded the plot to .png from the button and the yaxis is completely normal and not cut-off 😖
EDIT: Some of the downloaded png files are normal the rest show the cut-off y-axis

[cpsievert]

@timelyportfolio that's right, I'm drawing titles as annotations because x/y titles are always relative the plot in ggplot2, not each axis.

(in that comment, I think I meant to say "annotations aren't smart enough to dodge ticks & text"

[joshmyzie2]

Probably obvious, but an easy/crude workaround is to append "\n " to your x label and prepend "\n " to your y label. That should push your labels into the visible region.

[timelyportfolio]

Yes, but will worsen the situation when the plot is small. @cpsievert, do you have an example where the real plotly axis labels don't work?

[cpsievert]

qplot(data = mtcars, x = mpg, y = wt, facets = ~vs)

[drcrook1]

I'm having the same issues: http://stackoverflow.com/questions/36988379/adjust-axis-positions-ggplot2-facets. Also note that my titles are affected.

[timelyportfolio]

@cpsievert, if no facets, should we change the behavior to use plotly defaults. It seems we unintentionally prioritizing facet behavior over non-facet behavior.

[cpsievert]

@timelyportfolio yea, that's a good point, a pull request would be most welcome :)

[timelyportfolio]

@cpsievert I took a first stab at a pull request to make this change. I'm not sure that I got the axis title formatting correct. I tried to still use your margin calculations for both facetted and non-facetted plots. I know you will be much more familiar with the edge cases. I would love your help thinking through the best way to test.

[cpsievert]

This should now be fixed for non-faceted plots, but we'll need a more general solution for facets.

[nathaneastwood]

We seem to be seeing some of the plot being cut off (in our case the top of a bar chart) - possibly as a result of this - when used in Shiny. I can try to make a reproducible example if needed, but I don't have the source code. Interestingly the plot is not cut off when just used in the RStudio Viewer.

[sulaxd]

try setting layout(margin = list(b = 90))

[cpsievert]

Fixed via 9afbb5d

[ldecicco-USGS]

I was having these margin issues, so I tried installing via github to get this fix.

Now, I get an error:

library(ggplot2)
library(plotly)
df <- data.frame(x=1:10,y=1:10)
x2 <- ggplot(data = df) +
geom_point(aes(x=x,y=y))
ggplotly(x2)
Error in gg2list(p, width = width, height = height, tooltip = tooltip,  : 
  attempt to apply non-function

So, no problem, I'll just in-install this dev version of plotly and re-install via CRAN....but after tha, I was still getting that error.

I had a dev version of ggplot2 installed as well apparently (I was testing my code for the new updates that are coming out soon in ggplot2). Once I installed ggplot2 from CRAN, then that error finally went away.

I don't know if it's related to:
#233
but might be.

So, I still have the margin issues...but admittedly didn't install the fix...but there might be some other issue with the fix....just FYI