stat-quantile
Quantile regression statistic: fit one line per τ and sample it.
Returns one row per (τ, sample) with columns x, y, group (a per-τ key for grouping in line geoms), and _quantile (the τ value).
Usage
stat-quantile(
quantiles: (0.25, 0.5, 0.75),
n-samples: 64,
)Parameters
| Parameter | Default | Description |
|---|---|---|
quantiles |
(0.25, 0.5, 0.75) |
Array of τ values in (0, 1) to fit, e.g., (0.25, 0.5, 0.75). |
n-samples |
64 |
Number of evenly-spaced x positions sampled per fitted line. |
Returns
Statistic object with name: "quantile", consumed by geom-quantile.
Outputs
x.y.group._quantile.
Examples
Fit the 0.25, 0.5, 0.75 quantile lines through a noisy linear cloud via geom-line(stat: stat-quantile(...)); the stat’s group column auto-splits the lines.
#let n = 70
#let d = range(0, n).map(i => {
let x = i / n * 10
let noise = calc.sin(i * 0.7) + calc.cos(i * 1.9)
(x: x, y: 1.2 * x + noise)
})
#plot(
data: d,
mapping: aes(x: "x", y: "y"),
layers: (
geom-point(size: 1.5pt, alpha: 0.6),
geom-line(stat: stat-quantile(quantiles: (0.25, 0.5, 0.75)), stroke: 1pt),
),
width: 10cm,
height: 6cm,
)
