Changing Plot Font with Cairo in R

Question

. 。 Answers to this question are eligible for a +100 reputation bounty. teunbrand wants to to this question: ：

I'm confused about the same thing and I'd like an explanation about why this is the case and an answer giving an efficient way to get fonts to work with Cairo graphics (both png and pdf).

I have found R's default plots to be poorly aliased. As a solution, I set Cairo as the graphics device, and now the plots look much better.

Unfortunately, using Cairo has created another issue, which is that for some reason, I am not able to apply the font that I was using when the graph was displayed in the plot window (in the left-hand diagram above, Cambria is used, but the right-hand diagram fails to apply this font).

Here is my code:

library(readxl)
library(scales)
library(ggplot2)
library(dplyr)
library('Cairo')

windowsFonts(Cam = windowsFont("Cambria"))

dataset <- read_excel('CW Data.xlsx')
colnames(dataset)[4] <- "Broadband Subs (%)"

options(scipen = 1000)

# Scatter plot FDI~GDP with regression line
CairoWin()
ggplot(dataset, aes(x=`2019 GDP ($bn)`, y=`2019 FDI ($m)`)) + 
  geom_point(size=3, shape=1) +
  geom_smooth(method='lm',formula=y~x, se=FALSE, color='black') +
  scale_x_continuous(label = comma) + scale_y_continuous(label=comma) +
  theme(panel.background = element_rect(fill="peachpuff"), 
        plot.background = element_rect(fill="peachpuff")) +
  theme(panel.grid.major = element_line(colour = "gray72"), 
        panel.grid.minor = element_line(colour = "gray72")) + 
  theme(text = element_text(family = "Cam")) 

ggsave("FDI~GDP.png", device="png", type = "cairo") 

And here is a OneDrive link for the Excel data that I am using

https://1drv.ms/x/s!AvGKDeEV3LOs4gNr714Ie0KbOjhO?e=bkdPvk

Answer 1

Please don't consider this an answer to the question, but rather a motivating example for the bounty.

I found that the PDF device gives absolutely atrocious results using custom fonts. With regards to the bounty, here is a more minimal reprex.

library(ggplot2)

p <- ggplot(iris, aes(Sepal.Width, Sepal.Length, colour = Species)) +
  geom_point() +
  theme(text = element_text(family = "mono"))

ggsave("font_test.pdf", p, device = cairo_pdf)
#> Saving 7 x 5 in image

^{Created on 2022-02-04 by the reprex package (v2.0.1)}

A screenshot of the PDF output:

I needn't explain to the keen typographer that the displayed mono font is not 'Courier New' or system default monospaced font. Moreover, the left axis title is mangled and the legend title is out of whack.