Stacked area (ggplot2) and NA not considered

Question

I am using ggplot2 package to plot a stacked area of species abundance.

My code:

#Aggragate 2 column
    x=IBTS[,3:4]
    Datehour=paste(IBTS$Date.,IBTS$Time.)
    Date<-strptime(Datehour,"%d/%m/%Y %H:%M")

#Add a new column with the date
    IBTS$Date
    IBTS<-cbind(IBTS,Date)

#Plotting the data
    p<-ggplot(IBTS,aes(x=Date,y=Number.of.Particles.))
    p+geom_area(aes(fill=Selection.set.),position="stack")+  
    theme_bw()+ scale_fill_brewer(palette="Blues")

I obtain this plot:

I have a simply question: On Janv. 30, I have NA values on my dataframe but these values are not considered in my plot.

Does anyone have any idea?

Could you include a [reproducible example](http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example/5963610)? — Jaap, Feb 16 '16 at 11:49
what do you mean "not considered"? you would like to have a discontinuity (a white gap)? — agenis, Feb 16 '16 at 12:59
seems duplicate of http://stackoverflow.com/q/30116606/3871924 — agenis, Feb 16 '16 at 13:03
As Agenis asked, I would have discontinuity on this short period and would have the same result on the previous link. I tried with the geom_ribbon function but I have never built a loop... — Arnaud Louchart, Feb 16 '16 at 13:40
Could be the OP is complaining about a newish `ggplot2` bug then. Not sure what the problem is here, but if it is that he should have blanks (white space) where there is interpolation, then check this.http://stackoverflow.com/questions/35454277/can-you-make-geom-ribbon-leave-a-gap-for-missing-values/35461050#35461050 — Mike Wise, Feb 17 '16 at 19:54
I posted an issue and Hadley fixed it, so this might be fixed now. Not sure because it is `geom_area` and not `geom_ribbon`. https://github.com/hadley/ggplot2/commit/a760ff627582da59f938af14b0f01d54a4a97aa7 — Mike Wise, Feb 18 '16 at 02:25

score 2 · Answer 1 · edited May 23 '17 at 10:32

The complaint seems to be that there should be a blank white space at "Janv. 30" because there is missing data there. geom_area is based on GeomRibbon and there was a bug there that was causing this to happen which I posted to GitHub. Hadley actually just fixed it now, but the version is not released of course.

There is a workaround for now that I posted yesterday at this link Can you make geom_ribbon leave a gap for missing values?.

Update

Before I had you create a new geom to fix this, now I discovered a one-line that does the same thing:

GeomRibbon$handle_na <- function(data, params) { data }

p<-ggplot(IBTS,aes(x=Date,y=Number.of.Particles.))
p+geom_ribbon_na(aes(fill=Selection.set.),position="stack")+  
theme_bw()+ scale_fill_brewer(palette="Blues")

It is not an obvious dup since you have to know that the two layers geom_area and geom_ribbon use the same ggproto code

score 0 · Answer 2 · answered Mar 08 '16 at 09:28

0

I fixed the problem by creating 2 lines of null data according to the limits of the interpollation. Now, the new interpollation has made a "blank" a this place. I just want to thank you for your help

answered Mar 08 '16 at 09:28

Arnaud Louchart

61
6

Stacked area (ggplot2) and NA not considered

2 Answers2

Update