[ggplot2] Billiard ball "geom" with aestetics for factors - numbers in symbols

ggplot-billiard-ball-geom-w-aes-for-factors.ipynb

ggplot-billiard-ball-geom-w-aes-for-factors.R

library(ggplot2)
ggplot2::theme_set(theme_classic())

## Create a dataframe with dummy data
set.seed(123)
n <- 20
n_categories <- 5
df <- data.frame(
  x = 1:n + rnorm(n),
  y = 1:n + rnorm(n),
  category = sample(factor(letters[1:n_categories]), size = n, replace = TRUE))
summary(df)

## Create an initial graph
gg <- ggplot(data = df, aes(x = x, y = y, col = category, label = as.integer(category))) +
  geom_point(shape = 19, size = 6) +
  geom_text(col = "white", size = 3)
plot(gg)

## Problem
## -------
## How can we get numbers in the symbols in the figure legend?
##
## Solution
## --------
## The philosophy in the Grammar of Graphics and ggplot is that we map 
## abstract data to things we can perceive with our senses, aestetics.
## The solution to the problem is to think about the numbers as shapes 
## and not text, and then define your own mapping from category to shape.

## Define a mapping for the shapes on the billiard balls, i.e. the numbers on the balls
shapes_category <- as.character(1:length(levels(df$category)))
names(shapes_category) <- levels(df$category)
shapes_category

## Create a revised version of the graph
gg <- ggplot(data = df, aes(x = x, y = y)) +
  geom_point(aes(col = category), shape = 19, size = 6) +
  geom_point(aes(shape = category), col = "white", size = 3) +
  scale_shape_manual(values = shapes_category)
plot(gg)

## For fun, here's an alternative version
gg <- ggplot(data = df, aes(x = x, y = y)) +
  geom_point(aes(col = category), shape = 21, fill = "white", size = 6, stroke = 2) +
  geom_point(aes(shape = category), col = "black", size = 3) +
  scale_shape_manual(values = shapes_category)
plot(gg)