-
Notifications
You must be signed in to change notification settings - Fork 81
Open
Description
Given a dataset with two groups across time. We want to apply and annotate stats both independent (group1, time1) - (group1,time2) and paired (group1, time1) - (group1, time2)
From the examples it is not clear whether this can be done with one annotator?
statannotations.Annotator.Annotator has in its init a data and pair parameter, so the pairs need to be defined on initialization.
But apply/annotate, there we define the statistic.
So we need two Annotator instances and do like this?
annotator = Annotator(ax=ax, pairs=pairs_paired, plot="boxplot", data=data, x="condition", y="value", hue="group")
annotator.configure(test="t-test_paired") # within group, across time condition
annotator.apply_and_annotate()
annotator = Annotator(ax=ax, pairs=pairs_across_groups, plot="boxplot", data=data, x="condition", y="value", hue="group")
annotator.configure(test="t-test_ind")
annotator.apply_and_annotate()
However then the annotator instances do not know about each other and label can overlap, as shown below?
It could be fixed by forcing the stat across pairs (ttest-independent) to be on fixed y-value
#189
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
No labels