Just a lazy Monday afternoon
Sometimes, you and your significant other are attending a virtual statistics conference (eCOTS) and she decides to continue crocheting a Chicago flag-inspired baby blanket for her cousin’s baby shower (Ravelry “field of daisies” pattern linked here). Clearly, this is a common occurrence.
Figure 1: Chicago flag
She’s made 27 granny squares, all with a red circle. She wants to make a six square by nine square blanket, so needs 54 squares and isn’t confident on how to proceed, design-wise. Here are two exemplar squares:
And overall progress at this point:
You think that 54 red dots (every square) might not look that good, but are having a hard time making your case. Obviously, you need to create a ggplot2 so that she can compare the two pattern ideas!
The Mission
First, you need to find the colors of the Chicago flag:
Then, create a tibble (extension of data.frame) with the data needed to make the plot. inner* is the color of the center circle in the granny square, mid and outer are alternating colors, and x and y are the locations at which to plot the squares.
flag_colors = tibble(
inner1 = rep(red, 54),
mid = rep(c(lblue, white), 27),
outer = rep(c(white, lblue), 27),
x = rep(1:6, each = 9),
y = rep(1:9, times = 6),
# 2 - 4 not shown
inner2 = rep(c(red, lblue, white, red), times = 14)[1:54],
inner3 = rep(c(white, red, red, lblue), times = 14)[1:54],
picker4 = sample(rep(c(1, 2), times = 27)),
inner4 = if_else(picker4 == 1, outer, red),
# I learned mapping was "up and down"
# "cleverly" used matrix byrow parameter to swap to "left to right"
picker5 = rep(c(red, red, "a", "a"), times = 14)[1:54] |>
matrix(nrow = 9, ncol = 6, byrow = TRUE) |>
as.vector(),
inner5 = if_else(picker5 == "a", outer, red)
)Then, you write a function to plot any established inner* pattern. With a little help from Wolfram’s page on Regular Pentagons (equations 5 - 8), you are able to create a relatively high fidelity “daisy” shape.
patch_plot = function(data, inner, cons = 2.15, radius = 0.25){
c1 = 0.25*(sqrt(5) - 1)*radius
c2 = 0.25*(sqrt(5) + 1)*radius
s1 = 0.25*(sqrt(10 + 2*sqrt(5)))*radius
s2 = 0.25*(sqrt(10 - 2*sqrt(5)))*radius
ggplot(data = {{data}}, aes(x = x, y = y, color = mid)) +
geom_point(aes(color = outer), size = 16*cons, shape = 15) +
geom_point(aes(y = y + radius), size = 5*cons) +
geom_point(aes(x = x + s1, y = y + c1), size = 5*cons) +
geom_point(aes(x = x + s2, y = y - c2), size = 5*cons) +
geom_point(aes(x = x - s2, y = y - c2), size = 5*cons) +
geom_point(aes(x = x - s1, y = y + c1), size = 5*cons) +
geom_point(aes(color = {{inner}}), size = 4*cons) +
scale_color_identity() +
coord_fixed() +
xlim(c(0.5, 6.5)) +
ylim(c(0.5, 9.5)) +
theme_void()
}Finally, you want to compare two options – one with all red dots, and one with about half red dots.
amy = patch_plot(data = flag_colors, inner = inner1)
ryan = patch_plot(data = flag_colors, inner = inner5)
ggsave(filename = "amy.png", plot = amy, width = 7, height = 10)
ggsave(filename = "ryan.png", plot = ryan, width = 7, height = 10)
knitr::include_graphics("amy.png")
Figure 2: Amy’s preferred pattern
knitr::include_graphics("ryan.png")
Figure 3: Ryan’s preferred pattern
Conclusion
Alas, your significant other was not convinced. BUT you had fun making a ggplot function in the process, she was impressed with your ability to represent the design with ggplot, and she was more confident in her decision to complete the pattern as planned.