Exporting nice plots in R

from:http://www.r-bloggers.com/exporting-nice-plots-in-r/?utm_source=feedburner&utm_medium=email&utm_campaign=Feed%3A+RBloggers+%28R+bloggers%29

The Cairo-package

The Cario package is an excellent choice for getting antia-aliased plots in Windows. Here’s the same plot from the Cairo package (the background is now transparent):

^?View Code RSPLUS

1 2 3 4 5 6 7 8 9 10

library(Cairo) Cairo(file="Cairo_PNG_72_dpi.png", type="png", units="in", width=5, height=4, pointsize=12, dpi=72) my_sc_plot(data) dev.off()

If we want to increase the resolution of the plot we can’t just change the resolution parameter:

^?View Code RSPLUS

1 2 3 4 5 6 7 8 9

Cairo(file="Cairo_PNG_96_dpi.png", type="png", units="in", width=5, height=4, pointsize=12, dpi=96) my_sc_plot(data) dev.off()

We also have to change the point size, the formula is size * new resolution DPI / 72 DPI:

^?View Code RSPLUS

1 2 3 4 5 6 7 8 9

Cairo(file="Cairo_PNG_96_dpi_adj.png", type="png", units="in", width=5, height=4, pointsize=12*92/72, dpi=96) my_sc_plot(data) dev.off()

Adjusted point size results in the labels remaining the proper sizeIf we double the image size we also need to double the point size:

^?View Code RSPLUS

1 2 3 4 5 6 7 8 9

Cairo(file="Cairo_PNG_72_dpi_x_2.png", type="png", units="in", width=5*2, height=4*2, pointsize=12*2, dpi=72) my_sc_plot(data) dev.off()

 If you want the background to be white just add the bg parameter:

^?View Code RSPLUS

1 2 3 4 5 6 7 8 9 10

Cairo(file="Cairo_PNG_72_dpi_white.png", bg="white", type="png", units="in", width=5, height=4, pointsize=12, dpi=72) my_sc_plot(data) dev.off()

Here’s the plot code:

^?View Code RSPLUS

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

# Create the data data <- rnorm(100, sd=15)+1:100 # Create a simple scatterplot # with long labels to enhance # size comparison my_sc_plot <- function(data){ par(cex.lab=1.5, cex.main=2) plot(data, main="A simple scatterplot", xlab="A random variable plotted", ylab="Some rnorm value", col="steelblue") x <- 1:100 abline(lm(data~x), lwd=2) }