Powered by AppSignal & Oban Pro
Notesclub
Sign in with GitHub
created by hec & contributors
terms privacy

Logarithmic tick marks: guide_axis_logticks

notebooks/guide_axis_logticks.livemd

avatar
avatar
avatar
Share to X
Share to Bluesky
More notebooks

Logarithmic tick marks: guide_axis_logticks

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

About this notebook

This notebook follows the examples on ggplot2's reference page for guide_axis_logticks(), 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.

guide_axis_logticks() is an axis guide that places its tick marks at intervals in log10 space, in three lengths: long, middle and short. It suits an axis with a log scale or log-transformed coordinates; for data logged before plotting, it is told the base of that logarithm. Giggly has log scales, Giggly.Plot.scale_x_log10/2 and Giggly.Plot.scale_y_log10/2, but no axis guides, guide_axis_logticks() among them, and no coordinate system but Cartesian and radial.

Log scales

# A standard plot
p <- ggplot(msleep, aes(bodywt, brainwt)) +
  geom_point(na.rm = TRUE)

# The logticks axis works well with log scales
p + scale_x_log10(guide = "axis_logticks") +
  scale_y_log10(guide = "axis_logticks")

na.rm = TRUE only stops ggplot2 warning about the animals in msleep that have no brain weight; Giggly drops a row missing a position without a warning.

Not in Giggly yet: the msleep dataset, the na.rm option and guide_axis_logticks().

Log-transformed coordinates

# Or with log-transformed coordinates
p + coord_transform(x = "log10", y = "log10") +
  guides(x = "axis_logticks", y = "axis_logticks")

Not in Giggly yet: the msleep dataset, the na.rm option, coord_transform() and guide_axis_logticks().

Data logged by hand

# When data is transformed manually, one should provide `prescale.base`
# Keep in mind that this axis uses log10 space for placement, not log2
p + aes(x = log2(bodywt), y = log10(brainwt)) +
  guides(
    x = guide_axis_logticks(prescale.base = 2),
    y = guide_axis_logticks(prescale.base = 10)
  )

The mapping logs the data itself, base 2 for x and base 10 for y, and prescale.base tells each axis which base that was.

Not in Giggly yet: the msleep dataset, the na.rm option, and guide_axis_logticks() with its prescale.base option.

A pseudo-log scale

R's rcauchy() draws random numbers from the Cauchy distribution, whose tails reach far on both sides of 0, and set.seed(42) makes them the same on every run. A pseudo-log transformation is logarithmic far from 0 and close to linear near it, so it takes 0 and negative values too.

# A plot with both positive and negative extremes, pseudo-log transformed
set.seed(42)
p2 <- ggplot(data.frame(x = rcauchy(1000)), aes(x = x)) +
  geom_density() +
  scale_x_continuous(
    breaks = c(-10^(4:0), 0, 10^(0:4)),
    transform = "pseudo_log"
  )

# The log ticks are mirrored when 0 is included
p2 + guides(x = "axis_logticks")

Not in Giggly yet: a pseudo-log transformation, and guide_axis_logticks().

Ticks near zero

# To control the tick density around 0, one can set `negative.small`
p2 + guides(x = guide_axis_logticks(negative.small = 1))

Not in Giggly yet: a pseudo-log transformation, and guide_axis_logticks() with its negative.small option, the smallest absolute value that gets a tick.

Other notebooks:

Back