seaborn.objects.Plot can't realize bivariate distribution

[moonlighthalfwindow]

fig = plt.figure()
ax = fig.add_subplot(111)
p = (
    so.Plot(tips, x='tip', color='sex')
    .add(so.Bar(alpha=1), so.Hist(), so.Stack())
    .label(color='')
    .on(ax)
    .plot(pyplot=True)
)
p.show()

The y parameter passed in when Hist is called has no effect, it will have no effect on the graphing, it will simply generate a y-axis label and cannot generate a bivariate distribution.
I'm looking forward to the implementation of this feature.

[mwaskom]

Correct, this isn't implemented currently.

It would not be implemented as it is in sns.histplot because it doesn't make sense to portray a 2D histogram with a Bar mark. There would need to be (1) a Hist2D (or similar) Stat transform and (2) a suitable Mark, e.g. a heatmap mark. There's not currently a well defined way for Stat transforms to introduce a variable that is then mapped by one of the properties (i.e. the count would typically be mapped to the color; but you could also do something like use a Dot mark and map the count to dotsize).

Closing as not a bug.