Quantile regression: geom_quantile

Mix.install([
  {:giggly, path: Path.join(__DIR__, "..")},
  {:kino, "~> 0.19"}
  ])

About this notebook

This notebook follows the examples on ggplot2's reference page for geom_quantile(), which covers stat_quantile() too, in the order the page gives them. Each example shows ggplot2's code first and then Giggly's. An example Giggly cannot draw yet keeps its ggplot2 code and names what it is waiting for, so the page is also a list of what remains.

geom_quantile() fits a quantile regression of y on x and draws a line for each quantile it fits: a continuous analogue of a box plot. stat_quantile() does the fitting, with the quantreg package's rq(), or its rqss() for smooth quantiles. Giggly has none of this yet.

The quartiles and the median

1 / hwy turns miles per gallon into gallons per mile. Unless quantiles says otherwise, the lines are at the quantiles 0.25, 0.5 and 0.75.

m <-
  ggplot(mpg, aes(displ, 1 / hwy)) +
  geom_point()
m + geom_quantile()

Not in Giggly yet: geom_quantile().

Choosing the quantiles

m + geom_quantile(quantiles = 0.5)
q10 <- seq(0.05, 0.95, by = 0.05)
m + geom_quantile(quantiles = q10)

The first draws the median alone, the second a line at every 0.05 from 0.05 to 0.95.

Not in Giggly yet: geom_quantile() and its quantiles option.

Smooth quantiles

# You can also use rqss to fit smooth quantiles
m + geom_quantile(method = "rqss")
# Note that rqss doesn't pick a smoothing constant automatically, so
# you'll need to tweak lambda yourself
m + geom_quantile(method = "rqss", lambda = 0.1)

Not in Giggly yet: geom_quantile() with its method and lambda options and the smooth quantile regression it fits, and the quantreg package's rqss().

A fixed colour, width and transparency

# Set aesthetics to fixed value
m + geom_quantile(colour = "red", linewidth = 2, alpha = 0.5)

Not in Giggly yet: geom_quantile().