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
The tip on the linearRegression mark is also a bit chaotic (#1835), since it points at invisible dots. I don't see why someone would activate it on that mark, though, but an idea might be to display information about R2 (?). Though probably we should just drop the option? Doing nothing is probably fine too?
The text was updated successfully, but these errors were encountered:
I find the current behavior rather confusing. Doing nothing (and perhaps including a console warning?) for tip: true might be preferable. Even better options might include providing tips for "sample" points along the regression line, showing the current predictor value alongside predicted regression values and (if visible) +/- error values. The existing tip behavior for area marks is already pretty close to this. The R^2 value would also be nice, but probably not absolutely necessary.
A particular difficulty is that the regression is computed in screen space, which is equivalent (visually) for linear scales, and makes the mark work well with log scales etc; however it doesn't make it easy to derive predicted values. I also think we should do nothing and issue a warning — this would be better than the current behavior in any case. (And leave the door open if someone has a better idea in the future.)
The tip on the linearRegression mark is also a bit chaotic (#1835), since it points at invisible dots. I don't see why someone would activate it on that mark, though, but an idea might be to display information about R2 (?). Though probably we should just drop the option? Doing nothing is probably fine too?
The text was updated successfully, but these errors were encountered: