Apologies if this question is too simplistic, I feel like I see this type of chart all the time but I've never had to make one myself until now.

I'm teaching some non-mathematical people about how sample sizing works, and my example data look like this:

testid <- c(1, 1, 2, 2, 3, 3)
source <- rep(c("A", "B"), 3)
n <- c(1, 1, 100, 100, 100, 100)
p <- c(1, 0, 1, 2, 50, 51)

df <- data.frame(testid, source, n, p)
df
  testid source   n  p
1      1      A   1  1
2      1      B   1  0
3      2      A 100  1
4      2      B 100  2
5      3      A 100 50
6      3      B 100 51

What I would like to do is represent this with images that have n dots in a grid, of which p dots are filled in. So my output would look like:

{1A: image of one filled-in dot}
{1B: image of one empty dot}
{2A: image of 100 empty dots in a 10x10 grid, of which one is filled in}
{2B: image of 100 empty dots in a 10x10 grid, of which two are filled in}
{3A: image of 100 empty dots in a 10x10 grid, of which 50 are filled in}
{3B: image of 100 empty dots in a 10x10 grid, of which 51 are filled in}

Is this do-able in R (any package)? If not, what tool is best for doing it?

CodePudding user response：

Sure could this be achieved in R. As you want multiple charts I would write a function, which

1. creates a grid of n points via e.g. expand.grid
2. randomly chooses p grid points to fill
3. uses ggplot2 to plot the grid

Depending on the number of charts you could use this function to make your charts one by one. A more elegant solution would be to make use of e.g. purrr::pmap to loop over the rows of your dataframe to make your charts:

library(ggplot2)
library(purrr)

set.seed(123)

plot_grid <- function(n, p, ...) {
  # Make grid
  n <- round(sqrt(n))
  d <- expand.grid(x = seq(n), y = seq(n))
  d$fill <- 0
  # Randomly choose p grid points to fill
  d$fill[sample(seq(nrow(d)), p)] <- 1 
  # Plot
  ggplot(d, aes(x, y, fill = factor(fill)))  
    geom_point(shape = 21, size = 10)  
    scale_fill_manual(values = c("0" = "white", "1" = "red"))  
    theme_void()  
    guides(fill = "none")
}

grid_plots <- purrr::pmap(df, plot_grid)
names(grid_plots) <- paste0(df$testid, df$source)

grid_plots$`3B`