Plots.jl: Marginal distributions

[diegozea]

It would be great to have:

• [x] marginal histograms
• [ ] marginal density plots
• [ ] marginal boxplots
• [ ] marginal ticks

I imagine a xmarginal and ymarginal (marginal for both) keyword arguments taking: :boxplot, :histogram, :density, :ticks or a combination of those symbols (tuple).

[tbreloff]

These are all good ideas as recipes... I don't think they should be "built in" for Plots.

marginal histograms
marginal density plots
marginal boxplots

I need to allow a little more control over subplot layouts, but this should be done with linked subplots:

using Plots
pyplot(size=(500,400), grid=false)
n = 10000
x, y = randn(n), randn(n)

subplot(
    histogram(x),
    plot(),
    histogram2d(x, y, nbin=10),
    histogram(y, orientation=:h),
    link=true
)

marginal ticks

A nice implementation would require the ability to draw over the plot... this functionality will vary heavily by backend. I'm open to implementation ideas.

[tbreloff]

See #60

[tbreloff]

These are going in https://github.com/JuliaPlots/PlotRecipes.jl

[diegozea]

That's awesome! :)

[pkofod]

That is super cool! Yay for recipes

[tbreloff]

closing in favor of "Recipe ideas" issue