library(hexSticker)
library(ggplot2)
library(magick)Making Hex Stickers in R
Walk through any data science conference or browse the R community online and you will see them everywhere. The hex sticker has become one of the visual signatures of the R ecosystem. Most major packages have one, and researchers increasingly make them for their own projects, labs, and areas of work.
I have always liked the idea that something as technical as a research method can be turned into a small piece of visual identity. So, in this post, I am going to make a few hex stickers in R.
The stickers are made with the hexSticker package. The basic idea is to create a graphic, preferably with a transparent background, and pass it to sticker() along with the text, colors, and positioning you want.
A hex sticker has only a few elements to worry about. h_fill and h_color control the hexagon’s fill and border, while h_size controls the border thickness. The graphic placed inside the sticker is controlled through arguments such as s_x, s_y, s_width, and s_height. The package name or title is controlled by package, with p_x, p_y, p_size, p_color, and p_family controlling its appearance. You can also add a URL along the bottom with url. The sticker() function has many more options for controlling the appearance and layout of a sticker. You can read more about them in the hexSticker documentation.
My first sticker is for anthropoR, a small R package I wrote for computing BMI and classifying obesity. I already have a bmi_gauge.png graphic, which is a transparent, color coded arc that runs from low to high with a needle indicating the measurement. I’ll place this graphic inside the hexagon.
img <- image_read("bmi_gauge.png")
s1 <- sticker(
img,
package = "anthropoR",
p_size = 12, p_y = 0.65, p_color = "white",
p_fontface = "italic",
s_x = 1, s_y = 1.15, s_width = 5.8, s_height = 1.5,
h_fill = "#e63946",
h_color = "#7a1420",
h_size = 1.5,
u_size = 3.2, u_color = "white"
)
print(s1)
The next sticker represents missing data and imputation with a simple grid of data cells, some filled and others empty with dashed outlines to convey incomplete data.
set.seed(3)
grid <- expand.grid(row = 1:4, col = 1:4)
grid$missing <- runif(nrow(grid)) < 0.28
cells <- ggplot(grid, aes(col, -row)) +
geom_tile(data = subset(grid, !missing),
fill = "white", width = 0.8, height = 0.8) +
geom_tile(data = subset(grid, missing),
fill = NA, color = "#ffd166", linewidth = 0.8,
linetype = "dashed", width = 0.8, height = 0.8) +
theme_void() + theme_transparent()
s2 <- sticker(
cells,
package = "imputation",
p_size = 13, p_y = 0.45, p_color = "white",
s_x = 1, s_y = 1.05, s_width = 1.1, s_height = 1.1,
h_fill = "#8338ec",
h_color = "#4a1d8a",
h_size = 1.5
)
print(s2)
The last sticker uses the three curves from an SIR model for susceptible, infected, and recovered individuals over the course of an epidemic. I use a simulation from the deSolve package, so the epidemic shape shows a real model simulation.
library(deSolve)
library(tidyr)
sir <- function(time, state, parms) {
with(as.list(c(state, parms)), {
dS <- -beta * S * I
dI <- beta * S * I - gamma * I
dR <- gamma * I
list(c(dS, dI, dR))
})
}
out <- as.data.frame(ode(c(S = 0.99, I = 0.01, R = 0),
seq(0, 60, by = 1), sir,
c(beta = 0.4, gamma = 0.1)))
out_long <- pivot_longer(out, cols = c(S, I, R),
names_to = "compartment", values_to = "value")
curves <- ggplot(out_long, aes(time, value, color = compartment)) +
geom_line(linewidth = 1.1) +
scale_color_manual(values = c(S = "#e9f5f3", I = "#e76f51", R = "#ffd166")) +
theme_void() + theme_transparent() +
theme(legend.position = "none")
s3 <- sticker(
curves,
package = "SIRsimR",
p_size = 12, p_y = 0.5, p_color = "white",
s_x = 1, s_y = 1.05, s_width = 1.5, s_height = 0.95,
h_fill = "#2a9d8f",
h_color = "#1b5e57",
h_size = 1.5,
)
print(s3)
Hex stickers are genuinely fun part of the R ecosystem. Making one for your own research turns the visuals you already create into a piece of identity that you can use. The hexSticker package handles the geometry, leaving the enjoyable part to you. You get to choose the graphic and adjust the colors, fonts, and layout until everything looks right.
The designs here are just starting points. The same approach works for whatever your own research focuses on. Build the a transparent graphic, pass it to sticker(), keep the design simple and high contrast, and export it at print resolution.