Use of aes(color=category) in geom_density or geom_hist

[dmbates]

Describe how it works in R's ggplot2
A clear and concise description of what you want to happen.

If x is a numeric variable and g is a factor in data.frame df in R then ggplot(df, aes(x = x, color=g) + geom_density() produces a plot with multiple kernel density estimates. See the diamonds example in https://ggplot2.tidyverse.org/reference/geom_density.html

In TidierPlots.jl it seems that the color/colour aesthetic is interpreted as a scalar color by geom_density.

Is there a way to create a plot with multiple kernel density curves in TidierPlots.jl?

BTW, @rdboyes It is great to see someone from the Kingston area (are you on Wolfe Island?) in Julia development. I did both my undergraduate degree and my Ph.D. at Queens.

[rdboyes]

Yep! Wolfe Island - just finished my Ph.D at Queen's last year.

Thanks for reporting this - it's another issue downstream of the grouping issue (like #93) - looks like @adknudson has an idea how to fix it over there so we should have this working soon!

[rdboyes]

In dev branch now:

ggplot(penguins) +
         geom_density(aes(x = :bill_length_mm, color = :island))

[dmbates]

This seems to have reverted to the earlier behavior. Using TidierPlots v0.7.7 or the main branch of this repository, I get

[rdboyes]

Interesting - it seems to be plotting all three curves independently, but all in blue. I'll look into this change - it could be due to the new version of Makie? But I'm not sure yet

[rdboyes]

This is "re-fixed" in 0.8.0 (to be released soon)