You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I wanted to check how open the seaborn team would be to adding an additional param to the swarmplot() function to plot asymmetric swarmplots as well.
For example, users could potentially call
with a new param side: "r" | "l" | "c", which represents the sides "right", "left", "center" respectively that the swarm plot would be aligned against. ("center" would be the same as how the swarmplots are currently being plotted)
perhaps in more alignment with the API of the current violinplot() function, a new param split instead could be used to reproduce the same plot as the one above.
Several of the categorical functions produce asymmetrical plots that might make sense to draw "half" of this way. Would adding this just to swarmplot introduce new API inconsistencies?
While you mention that violinplot has a similar feature, it's not identical: a single split violin (as of v0.13.0 where this became possible) spans the entire width of the space allotted for a non-split violin. Maybe that is actually not ideal and it would be better for odd splits to occupy half the space, but that does mean that just adding split to swarmplot to make the plot you're showing above doesn't make much sense.
What happens with hue is used? Would you have multiple colors on swarms with the shapes your'e showing above? People would probably want, if not expect, that the swarm might "split" the way the violin plot does (they certainly would if the split parameter is used).
Hi!
I wanted to check how open the seaborn team would be to adding an additional param to the
swarmplot()
function to plot asymmetric swarmplots as well.For example, users could potentially call
side
: "r" | "l" | "c", which represents the sides "right", "left", "center" respectively that the swarm plot would be aligned against. ("center" would be the same as how the swarmplots are currently being plotted)violinplot()
function, a new paramsplit
instead could be used to reproduce the same plot as the one above.Both of which would result in the following plot below:
(The plot above has been produced via code edits I've made to the current
swarmplot()
function wrt point 1.)To sum it up,
Thank you so much!
The text was updated successfully, but these errors were encountered: