r - How to plot alphabets? - r

Question

I want a scatter plot which looks like letters of the alphabet. How can I do this with a program? I can just enter co-ordinates and make the plot look like an 'A' or 'S' or whatever. But can it be done in an easier manner?

Answer 1

The pch argument of plot will take arguments that can be used to represent these values. From ?points, values 32-127 are the ASCII character set.
With a little messing around, values 65:90 correspond to capital letters, and values 97:122 correspond to lower case letters.
To illustrate this, try
plot(1:10, 1:10, type="p", pch=97:107)
for example.
Here is a plot of all of the latin alphabet
# blank canvas
plot(1:30, 1:30, type="n")
# upper case
points(1:26, 1:26, pch=65:90)
# lower case
points(1:26, 4:29, pch=97:107)
You could even build a mapping between these values for easier reference.
myRefUpper <- setNames(65:90, LETTERS)
myRefUpper
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90
myRefLower <- setNames(97:107, letters)
myRefLower
a b c d e f g h i j k l m n o p q r s t u v w x y z
97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122
This way, you could refer to specific letters by name. For example, try
plot(1:10, 1:10, type="p", pch=c(myRefLower[c("a", "t", "q")], myRefUpper[LETTERS[10:16]]))

Answer 2

There is now an R package on GitHub that provide coordinates for the Hershey fonts that Ben Bolker mentioned: hershey.
For example, we can get coordinates for the start and end of each stroke (line) in the letter A, for the Roman Simplex font (a simple font using minimal straight lines to create letters):
library(hershey)
coord <- subset(hershey, font == 'rowmans' & char == 'A')
coord
#> x y left right width stroke idx glyph font ascii char
#> 93723 0 12 -9 9 18 0 1 34 rowmans 65 A
#> 93724 -8 -9 -9 9 18 0 2 34 rowmans 65 A
#> 93725 0 12 -9 9 18 1 3 34 rowmans 65 A
#> 93726 8 -9 -9 9 18 1 4 34 rowmans 65 A
#> 93727 -5 -2 -9 9 18 2 5 34 rowmans 65 A
#> 93728 5 -2 -9 9 18 2 6 34 rowmans 65 A
We can use the base approx function to interpolate between the start and end points for each stroke, then plot the result, using the graphical parameter pty to set a square aspect ratio:
op <- par(pty = "s")
plot(coord[, 1:2], type = "n")
for (i in unique(coord$stroke)){
points(approx(subset(coord, stroke == i)))
}
To reset the default graphical parameters:
par(op)
For ggplot2 you can do the interpolation first as below:
library(dplyr)
library(ggplot2)
coord2 <- coord %>%
group_by(stroke) %>%
do(as_tibble(approx(.)))
ggplot(coord2, aes(x, y, group = stroke)) +
geom_point() +
coord_equal() +
theme_minimal()
Created on 2022-01-05 by the reprex package (v2.0.1)
Edit
approx won't work for letters with strokes that have x values with different y values, e.g. vertical strokes or strokes that bend back on themselves. For this we can define our own linear interpolation function:
interp <- function(coord, eps = 0.5) {
y <- coord$y
x <- coord$x
n <- length(x)
x2 <- (x[-1] - x[-n])/eps
y2 <- (y[-1] - y[-n])/eps
p <- pmax(abs(x2), abs(y2))
id <- sequence(p)
list(x = c(x[1], rep(x[-n], p) + rep((x[-1] - x[-n])/p, p)*id),
y = c(y[1], rep(y[-n], p) + rep((y[-1] - y[-n])/p, p)*id))
}
library(hershey)
coord <- subset(hershey, font == 'rowmans' & char == 'C')
op <- par(pty = "s")
plot(coord[, 1:2], type = "n")
for (i in unique(coord$stroke)){
points(interp(subset(coord, stroke == i)))
}
Created on 2022-01-05 by the reprex package (v2.0.1)