Closed danicahelb closed 2 years ago
i commented on veupathdb/edanewissues#238 but looks like i should have here.. i was wondering if it might be a clearer ux for low cardinality continuous vars to only offer mean and median. is there a really strong use case for offering the proportion option for them given we can now forceStringType
if needed and appropriate?
I don't have a strong use case for allowing proportion in addition to mean & median. However, in cases of continuous variables with <10 values, it would make sense to also offer proportions for flexibility for future use cases I haven't yet considered. we'll discuss at a future EDA mtg
The help text would need changing if we offered the choice of mean/median/proportion for these variables, and that would NOT do! :rofl:
I've just created a THIRD ticket for this issue: https://github.com/VEuPathDB/web-eda/issues/1066
I'll consolidate them somehow... sorry.
Duplicate of https://github.com/VEuPathDB/web-eda/issues/1066 will be closed (for b57 at least) by https://github.com/VEuPathDB/web-eda/pull/1067
Here is a use case from GEMS1. "Household wealth index, categorical" are actual categories (low --> high), but they are represented by numbers. I can see wanting to make a proportion here.
Though perhaps it would be better to update the value map in this case
Continuous terms with <10 unique values can only be plotted as proportions in the line plot. For these, it would be nice to enable the user to choose between all 3 options (mean, median & proportion)