Release 1.0.0

[strengejacke]

@DominiqueMakowski I think for version 1.0.0, most/all of these should be fixed, if possible.

Originally posted by @strengejacke in #287 (comment)

Tasks

Preview Give feedback

No tasks being tracked yet.

[DominiqueMakowski]

For plotting, currently w do a lot of guesswork to pick what best goes in x / color / alpha depending on the type of the variables (factor, continuous etc). This is probably over-complex and not optimal anyways, so we could simplify the rules and streamline the inner workings. We could have the same internal for every estimate_* functions with the following aesthetics:

• y = whatever is predicted, means, prediction, contrasts, slopes etc.
• x = first argument of by
• color = second argument of by
• alpha = third argument of by
• Everything else goes into facet_wrap

[strengejacke]

Sounds good! x could be either first value of by, or first value of contrast or trend (if available - I think we save all the information as attributes).

[strengejacke]

I think we should have a 0.9.0 release soon, we have added a lot of new features the last time. Which issues would you like to be fixed for 0.9.0?

[DominiqueMakowski]

We can focus on bugs / broken vignettes since there's not a lot and leave plotting refactoring for 1.0 ?

[DominiqueMakowski]

true, comparison sounds good

[DominiqueMakowski]

of type which is more generic

[strengejacke]

type could be confused with the type argument in predict(). Furthermore, comparison = "(b3-b1)=(b4-b2)" indicates a comparison, but not really a type? In the particular case of contrasts, comparison seems more generic to me.

[DominiqueMakowski]

agreed!

[strengejacke]

Just FYI, see docs for hypothesis here: https://marginaleffects.com/man/r/comparisons.html
and here: https://marginaleffects.com/chapters/hypothesis.html

[strengejacke]

Test coverage increased from ~36% up to 80%

[DominiqueMakowski]