How to remove the white margin of ggord in ggplot2? - r

I have this plot. I am using the following package and script to do a biplot. But the figure output always has white margins on the top and bottom of the actual figure. I tried removing them using plot.margin=unit(c(0,0,0, 0), but no luck. Any thoughts on how to do this or is it just the default and there is nothing that can be done?
The script (I am using the example, but the same issue with mine):
>library(devtools)
>library(ggord)
>ord <- prcomp(iris[, 1:4])
>p <- ggord(ord, iris$Species)
>p
The problem is the white margins on the top and bottom as seen here. How can I remove them?

The white space you are seeing is a function of the window size when you save the image. If you resize your window vertically or horizontally, you'll notice that the white space shrinks or grows.
If your goal is to save the plot as a PNG image, you can export it directly to file with your desired dimensions as follows:
png("plot1.png", 800, 300);
p;
dev.off();
Adjust the 800 to the desired length and the 300 to the desired height, in pixels. If the size is larger than the actual graph, it will pad with blank space.

Related

How to remove image margins when plotting and exporting square image with R?

I'm trying to plot a square TIFF image (1024 x 1024), but it displays it as a vertically stretched rectangle. I've tried adjusting margins in plotRGB and utilizing par(mar=c(0,0,0,0)), but I can't get the plot and output to be a square even though R recognizes the raster as a square.
Does anyone have any advice on how to simply fix this issue? I've attached two images, one of the original image and one of the image when plotted/exported in R. The goal of my program is to use histmatch solely to adjust the saturation of the source image to that of a matching reference image. The exported image thus cannot be vertically stretched, and must have no margin.
Screenshot of original image
R display and export of image

Using data line widths in bokeh

Is there a way to provide line width in bokeh in data space, or alternatively can the line width be fixed in a way that zooming does not alter it?
I am trying to draw lines right next to one another without clashes or gaps between them.
Currently, I am using a mapping from line width to coordinate width and change the line widths and line coords in sync to achieve this, but as soon as I zoom the line width is adapted.
Is there a way to provide line width in bokeh in data space
No, but feature requests are welcome!
can the line width be fixed in a way that zooming does not alter it?
Line width is set in screen units - zooming does not alter it by default.
It seems like the problem is that you change your line width in response to the range changes, and zooming does change ranges.

remove white space from table png

I am using the following code to export some tables as pictures. Using the code below can export the table but there is a lot of white space around it which I want to remove. Also is there a way to autosize the images as tables have varying rows?
png("picture.png", width = 650, height = 600, bg = "transparent")
grid.table(df)
dev.off()
Expected output will remove the white space that comes around the table and the actual picture should fit the size of the table only

R ggplot, remove white margins in ggsave/ggplot

How can I remove the white margins in ggsave?
My question is exactly the same as Remove white space (i.e., margins) ggplot2 in R. However, the answer there isn't ideal for me. Instead of trial and error for a fixed but unknown aspect ratio, I would like to give ggsave a height and weight and want my plot (ie top of title to bottom of x-label) to automatically expand to that configuration without white margin.
How can I remove the strange white margin around my .png (plotted with r, ggplot)? gives a way to make the margin transparent, but they are still there and the plot is smaller than height and width I set in the saved file.
Found the answer from Remove Plot Margins in ggplot2
theme(plot.margin=grid::unit(c(0,0,0,0), "mm"))
does the job
In this answer linking to this blog post there is a solution which also works for different aspect ratios. You can crop the image on your hard drive, independently of OS:
knitr::plot_crop()
If you're using Unix or Mac OS, another option when the various margin options aren't trimming enough is to use the pdfcrop command available within Unix through R's ability to invoke system commands:
# after saving image, run pdfcrop
system2(command = "pdfcrop",
args = c("name_or_path_of_file_before_crop.pdf",
"name_or_path_of_file_after_crop.pdf")
)
For more, see: https://robjhyndman.com/hyndsight/crop-r-figures/
If pdf and pdfcrop aren't your thing, for example you work in png with a png logo - then see my answer here: How to save a ggplot2 graphic with the proper aspect ratio?
I ended up adding a command like this after ggsave:
system("/usr/local/bin/mogrify -trim -border 8 -bordercolor white output.png")
-trim removes an existing margin and -border 8 -bordercolor white adds a small 8px margin around the plot.
For a plot that had a gray background, a few white pixels were left around the edges of the plot, so I used the -shave option to remove a few extra pixels:
system("/usr/local/bin/mogrify -trim -shave 4x4 output.png")

Set width and height of graphic made using ggplot, grid, and gridExtra

If I have a graphic composed of several plots, say three plots arranged vertically. This is a gtable object and can be drawn to the page with:
grid::grid.newpage()
grid::grid.draw(plot)
However I see that the plot in my RStudio is 'smushed up' as in the screenshot below:
As you can see in the bottom right corner it is squashed and the titles overlap with other elements of the graphic.
If I hit zoom and view the plot it is a lot bigger:
Now I know, that if I were to export my gtable plot using pdf() or png() and such devices, I can set a width and a height, and so just make it big enough such that the plot is not squashed.
However, instead of one of those graphic devices, I would like to use export.grid, from the gridSVG package to save it to an SVG file. But if I do
gridSVG::export.grid(plot)
Then the SVG file exported looks squashed as it does in the RStudio plot window.
So my question is, how can I manipulate the dimensions of the graphic so it is drawn to SVG without it looking squashed? I draw the plot initially with grid.newpage and grid.draw, I wonder perhaps I have to specify some size of the page or drawing using grid.
Thanks,
Ben.

Resources