Step 1: Understand the problem

The problem is with using scale_fill_stepsn in ggplot2 to color points based on a continuous variable. The issue is that the breaks are not set correctly, causing the colors to blend or interpolate.

Step 2: Identify the solution

To fix the issue, we need to set the breaks to be at the minimum and maximum values of the data, and use 8 breaks (the length of the palette + 1). We also need to map the colors to the center of each bin using the values argument.

Step 3: Provide a working solution

To achieve this, we can modify the code as follows:

ggplot(data) +
  geom_point(aes(x = x, y = y, fill = col), size = 10, pch = 21) +
  scale_fill_stepsn(colors = pal, 
                    breaks = seq(min(data$col), max(data$col), length = 8),
                    values = seq(1/14, 13/14, length = 7))

Alternatively, we can use helper functions to calculate the correct values:

get_breaks <- function(vec, breaks) {
  c(min(vec), breaks, max(vec))
}

get_values <- function(vec, breaks) {
  breaks <- get_breaks(vec, breaks)
  vals <- (diff(breaks)/2 + head(breaks, -1) - min(breaks))/diff(range(vec))
  vals[length(vals)] <- vals[length(vals)] + .Machine$double.eps
  vals
}

ggplot(data) +
  geom_point(aes(x = x, y = y, fill = col), size = 10, pch = 21) +
  scale_fill_stepsn(colors = pal, 
                    breaks = get_breaks(data$col, 1:6 * 0.01),
                    values = get_values(data$col, 1:6 * 0.01))

Step 4: Test the solution

The solution should fix the issue and produce a plot with correctly colored points.

The final answer is: There is no specific numerical answer to this problem as it involves providing a working solution to a ggplot2 issue.

Last modified on 2023-12-16