search

sefffal / PairPlots.jl

Beautiful and flexible vizualizations of high dimensional data
https://sefffal.github.io/PairPlots.jl/dev
MIT License
130 stars 7 forks source link

Column vs Row samples #40

Closed alecloudenback closed 7 months ago

alecloudenback commented 7 months ago

The docs say:

# As always in Julia, columns are treated as variables, and rows as samples.

but is that true? Using Distrubtions.jl, this will produce a 2x10000 matrix:

rand(MvNormal([0.,0.],[1 0.5; 0.5 1]),10000)

Copulas.jl also produces a wide matrix instead of long when sampling.

I'm not sure what's the preferred approach but I got tripped up here and found myself doing a number of transposes while experimenting with PairPlots.jl

sefffal commented 7 months ago

Hi @alecloudenback thanks for the report. I’m not keen to change this (ultimately we have to pick something!) but I am open to docs fixes. Do you have any wording suggestions that might have helped you the first time you used the package?

BTW this is more of a fallback. The primary way to use PairPlots is via a Tables.jl compatible table.

alecloudenback commented 7 months ago

Maybe just drop the "As always in Julia" part?

sefffal commented 7 months ago

Done. Thanks for the feedback!