Why are Groups with 'NA'-Values dropped in pirateplot?

[tinu-schneider]

I have data that has no values for group B (all NA). If I make a pirateplot, the group B is not plottet.

   Group     Year   Value
1      A     2010      10
2      A     2015      15
3      B     2010      NA
4      B     2015      NA
5      C     2010      20
6      C     2015      25

Is there a possibility that I can include group B? In my case I want to show, that we got no results for group B. And I prefer keeping the values as NA and not changing them to 0.

Here is a minimal example

dat <- data.frame(
       Group = c('A', 'A', 'B', 'B', 'C', 'C'),
        Year  = rep(c('2010', '2015'), 3),
        Value = c(10, 15, NA, NA, 20, 25)
)
yarrr::pirateplot(formula = Value ~  Year + Group, data = dat)

[DominiqueMakowski]

Hi!
Not sure if it answers your question, but I believe that it is better that yarrr works this way. NA values are not equal to 0 (missing data could, in theory, take any possible value) and usually, we don't want them to be treated as 0's... So in your case if your NA's are actual 0's, I think it would be simpler to change them accordingly...

Cheers 😉

[tinu-schneider]

Ok - I see. So I tweak the plotting on my own like this.

• If all values of a group are NA, set one of them to the min value of all other groups
• plot a white polygon over the added faked values (here group B)
• add NA to the plot
  

As I mentioned before: I would like to show the fact, that we didn't get any data for group B, in our case this is an important fact that should not be dropped.

Thanks anyway!

[ndphillips]

@tinu-schneider Yes I see how visualising missing data could be important in some cases (though in others it might be a nuisance...). I'll work on adding a new logical argument called ignore.missing that controls this.

[tinu-schneider]

Sounds good!