cornerplot and corrplot

[juliohm]

Any plans to add cornerplot and corrplot like in StatsPlots.jl? Perhaps it is already possible to achive similar result with AOG?

[piever]

Something like

df = (x = rand(100), y = rand(100), z = rand(100))
vars = [:x, :y, :z]
data(df) * (visual(Scatter) + linear()) * mapping([:x, :y, :z], [:x :y :z], col = dims(1), row = dims(2)) 

would do a "symmetric" corrplot. It's definitely planned to have the full corrplot functionality available, but the interface / implementation are not super clear. See also #187 and piever/SplitApplyPlot.jl#53 (comment)

[juliohm]

@piever can you please clarify the col/row mechanism? Couldn't find it in the docs. I am having trouble creating the corner/corrplot using the snippet you shared above.

[filchristou]

I am also having some trouble undestanding how does this col, row, dims mechanism work.
In your example your diagonal will be just a straight line.

is there a way to substitute the diagonal plots with the density functions instead ?

[sethaxen]

Something like

df = (x = rand(100), y = rand(100), z = rand(100))
vars = [:x, :y, :z]
data(df) * (visual(Scatter) + linear()) * mapping([:x, :y, :z], [:x :y :z], col = dims(1), row = dims(2)) 

would do a "symmetric" corrplot.

@piever I have data in x and would like to split x by a categorical cycling variable group, plotting a 4x4 corner plot where subplot (i, j) compares x-values of group i with x-values of group j. Can this example be modified for something like this? e.g. data

df = (x = rand(100), group=repeat(1:4, 25))

[piever]

I think something like

data(df) * (visual(Scatter) + linear()) * mapping(:x, :x, col = :group => nonnumeric, row = :group => nonnumeric)

should do the trick in that case.

EDIT: as pointed out below, this does not do the desired thing.

The things that are not supported at the moment are adding marginal histograms in a simple way (either around the layout or in the diagonal) and doing asymmetric things (lower triangle is different from upper triangle). I'm not sure if there is a simple way to support that consistently within AoG, or whether one is better off adding some sort of "recipe with rich layout" to Makie, where a plot gets a subfigure instead of just an axis. There is a little bit of discussion on that around piever/SplitApplyPlot.jl#53 (comment)

[sethaxen]

I think something like

data(df) * (visual(Scatter) + linear()) * mapping(:x, :x, col = :group => nonnumeric, row = :group => nonnumeric)

should do the trick in that case.

This doesn't seem to:

[piever]

Ah yes, of course you are right... I'm afraid there is no way at the moment other than manually unstacking the data to pass to the wide format.

[pkofod]

Is this something you would be interested in adding as a feature? And have you already thought about the "AoG" interface for such a thing? Or do you think it should just be a Makie recipe with DataFrame input because it can't really leverage any AoG features?

[juliohm]

Update: PairPlots.jl provides this functionality with Makie.jl Em ter., 19 de set. de 2023 03:06, Patrick Kofod Mogensen < ***@***.***> escreveu:

…

Is this something you would be interested in adding as a feature? And have you already thought about the "AoG" interface for such a thing? Or do you think it should just be a Makie recipe with DataFrame input because it can't really leverage any AoG features? — Reply to this email directly, view it on GitHub <#210 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAZQW3I6GPSOHRJXCFFEUZ3X3EY67ANCNFSM46L756PQ> . You are receiving this because you authored the thread.Message ID: ***@***.***>