set.seed(1)
n <- 1e6
df <- data.frame(x = rnorm(n), y = rnorm(n) + rnorm(n, 0, 0.4))
p <- vplot(df) |>
mark_datashade(x = x, y = y) |>
labs(title = "One million points, datashaded")Datashading a million points
Big data
Aggregate-then-shade rendering that stays fast and legible at a million points.
A scatter of a million points is just ink. mark_datashade() uses vellum’s built-in datashader-style fast path: aggregate points into a pixel grid, then shade by density, so the picture scales with the canvas, not the data.
