Could the x and y axis position move when `coord_radial` is used?

[davidhodge931]

Currently only the y axis title moves

library(tidyverse)

mpg |> 
  ggplot() +
  geom_boxplot(aes(x = class, y = displ)) +
  coord_radial() +
  scale_y_continuous(position = "right")

^{Created on 2024-03-02 with reprex v2.1.0}

[teunbrand]

It is a little bit more complicated than just moving the axis.
In particular, if you have start/end settings that don't include a vertical angle, the axis is placed horizontally.
So position = "right", or position = "left" for that matter, cannot apply correctly.

library(ggplot2)

mpg |> 
  ggplot() +
  geom_boxplot(aes(x = class, y = displ)) +
  coord_radial(start = 0.25 * pi, end = 0.75 * pi, r_axis_inside = FALSE)

Also if you have a bottom-half circle, the position would be on the right visually even though it is the primary axis.

mpg |>
  ggplot() +
  geom_boxplot(aes(x = class, y = displ)) +
  coord_radial(start = 0.5 * pi, end = 1.5 * pi)

If you must have a guide on the right, you can use the r.sec name in guides().

mpg |> 
  ggplot() +
  geom_boxplot(aes(x = class, y = displ)) +
  coord_radial() +
  guides(r = "none", r.sec = "axis")

^{Created on 2024-03-04 with reprex v2.1.0}

[davidhodge931]

It works for coord_polar, so it'd be cool if could work for coord_radial. But I see it's a bit more complicated than I'd thought.

An idea is to just switch by 180 degrees where-ever the axis was, if someone uses position = "right" or "top"? That would make sense when it was a full or half pie, which would cover off most situations.

Feel free to close

[teunbrand]

There is something to say for this from a consistency standpoint. However, I don't think it makes sense to use a flipping mechanism for the theta axis: that would only work if inner.radius allows enough space.