change image size rmarkdown presentation

Change image size rmarkdown presentation full#
Change image size rmarkdown presentation download#

Custom fonts may not render correctly on other systems. The look of the output may change depending on which program opens them. There are valid reasons to not want to use vector graphics, however. With these you can simply render the graphic to a size where everything looks as it should and then resize the output to the physical size that you need. One way to resolve this is to not render to raster but use a vector graphic device such as The discussion above is especially relevant to raster output since they don’t resize gracefully and it is thus very important to get the correct dimensions and resolution when it is rendered. Now, how should we go about correcting this? Attempt 1: Use vector graphics If you render a plot to a smaller size than the one you’ve previewed you may find that text, margins, etc.

Change image size rmarkdown presentation full#

This means that as we scale up the output size, they remain the same size and will thus get smaller relative to the full image. The same is true for the element sizes in the scatterplot, the grid lines, etc. The text is given in points which, as you recall, is an absolute dimension. While the relative sizes expand along with the output size, the absolute sizes does not. Why is that? Our plot is a mix of elements positioned and dimensioned based on both relative and absolute sizes.

The downscaling done by the webpage allows us to simulate how looking at our poster may feel like, and it is not pretty everything seems super small and ineligible.

Change image size rmarkdown presentation download#

You can download each of the images and open them in a image editor to convince yourself that they are of different absolute size. The plot above doesn’t look physically larger, but that is because the webpage downscales it to make it fit. Pngfile <- fs:: path( knitr:: fig_path(), 'large_basis.png')Īgg_png( pngfile, width = 60, height = 36, units = 'cm', res = 300) With that in mind we quickly size up our plot: How much bigger? Well, if we assume that we are watching our screen at 50 cm distance, and our poster is meant to be observed at 1.5 m distance, then our plot should be 3 times larger to take up the same amount of space in our vision: A poster is usually observed at a farther distance than a computer screen, so in order to make it legible the plot should be bigger. This looks good, but remember we want to use this on a poster. # I'm explicitly calling the device functions so you can see the dimensions # used agg_png( pngfile, width = 20, height = 12, units = 'cm', res = 300) Pngfile <- fs:: path( knitr:: fig_path(), 'basis.png') ) + labs(x = 'Flipper Length ', y = 'Body Mass ', colour = 'Species', # plot adapted from p <- ggplot( na.omit( penguins), aes(x = flipper_length_mm, y = body_mass_g)) + geom_point(ĭescription = 'Gentoo penguins are generally bigger in size') This is the plot, and how it looks on a computer: In the remainder of the text we will assume that we sit in front of a computer developing a plot that should end up on a poster. Once we are content with it we will save it to the correct absolute size required by wherever we intend to publish the plot. When we develop a graphic we will generally sit in front of a computer and fine tune it while continuously getting previews on the screen. At its core, this is about ensuring the correct scaling of a plot as we develop it for varying absolute sizes. With formalities out of the way, we can describe the problem more clearly. Since points is an absolute unit, the resolution of the output will determine the number of pixels it correspond to. Varied throughout history, the general consensus now is that 1pt = 1/72 inch (this is also adopted by R). When we set a font to size 12, it is given in points. Pointsize: This is a measure tied to text sizing. A resolution of 72 ppi means that an inch is considered 72 pixels long. It is usually given in ppi (pixels per inch), though dpi (dots per inch) is used interchangeably. Resolution: This number ties absolute and pixel size together. Pixels does not have any inherent physical size. The pixel size is the number of rows and columns in the matrix. Pixel size: For raster output, the graphic is encoded as a matrix of color values. This is measured in centimeters or inches or another absolute length unit. Some definitionsīefore we delve into the problem we should clarify a few concepts related to graphics and sizing:Ībsolute size: This is the physical dimensions of the graphic (or, more precisely, the intended physical dimensions). The latest release of ragg contains a new functionality that will hopefully make this issue a thing of the past.

By a very large margin the most “popular” response revolved around making sure that output had the correct scaling of text, lines, etc. Some time ago, while working on the new edition of the ggplot2 book, I asked out to the R twitterverse what part of using ggplot2 was the most incomprehensible for seasoned users.