Include number of missing values in ggplot

Question

I'm using ggplot2 to plot different time series (one for Alice, one for Bob, one for Eve), which have a different number of missing values.

require('ggplot2')
df3 <-  data.frame(name=c(rep("Alice",10),rep("Bob",10),rep("Eve",10)),value=c(seq(1,10), seq(4,13), seq(5,14)), time=rep(seq(1,10),3))
df3$value[c(3,4,15,16,17,22,23,24,25)]<- NA
ggplot(data=df3, aes(time, value)) + 
  geom_line() + 
  geom_point() + facet_wrap(~ name, nrow=1)

I'd like to have the count of NAs displayed in each of the plots, e.g. as an overlay of a number (2 for Alice, 3 for Bob, 4 for Eve). Is there an elegant way to do this?

create a dataset with the number of NA for each name and then use geom_text() to add the values to the plot. — MLavoie
– MLavoie, Commented Jan 8, 2016 at 20:24
I asked a similar question and got help here. stackoverflow.com/questions/25081619/… — Reuben Mathew
– Reuben Mathew, Commented Jan 8, 2016 at 20:43

boshek · Accepted Answer · 2016-01-09 20:43:05Z

3

As @MLavoie suggested in the comments, generate a new dataframe for the text labels then work with that. This should work for your purposes:

require('ggplot2')
require('dplyr')

df3 <-      data.frame(name=c(rep("Alice",10),rep("Bob",10),rep("Eve",10)),value=c(seq(1,10), seq(4,13), seq(5,14)), time=rep(seq(1,10),3))
df3$value[c(3,4,15,16,17,22,23,24,25)]<- NA

NAdf<-df3 %>%
  group_by(name) %>%
  summarise(ycoor=mean(value, na.rm=TRUE),
            xcoor=mean(time, na.rm=TRUE),
            num_NA=sum(is.na(value)))  


ggplot(data=df3, aes(time, value)) + 
 geom_line() + 
 geom_point() + 
 geom_text(data=NAdf, aes(x=xcoor, y=ycoor, label=paste(num_NA,"for",name))) +
 facet_wrap(~ name, nrow=1)

HTH

Updated

In response to the comment below. Generally I find placing text labels into a facetted plot fairly finicky. In your example you could simply define the x and y coordinates as 5,5 for all panels like this:

NAdf<-df3 %>%
  group_by(name) %>%
  summarise(ycoor=5,
            xcoor=5,
            num_NA=sum(is.na(value)))

Then you could plot using the same code as before:

 ggplot(data=df3, aes(time, value)) + 
 geom_line() + 
 geom_point() + 
 geom_text(data=NAdf, aes(x=xcoor, y=ycoor, label=paste(num_NA,"for",name))) +
 facet_wrap(~ name, nrow=1)

The issue with this is that it isn't a generalized solution. In practice though I find you need to fiddle with your geom_text plotting coordinates each and every time to get it just right. Truth be told @Sam Dickson's solution is very elegant for this particular problem.

edited Jan 9, 2016 at 20:43

answered Jan 8, 2016 at 20:48

boshek

4,4462 gold badges37 silver badges68 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Roland Over a year ago

What's the %>%? I'm getting an error: Error: could not find function "%>%"

boshek Over a year ago

You need to load dplyr. Also it is %>% not %<%. Think of %>% (called a pipe) as the word "then". It is automatic for me to use pipes these days.

Roland Over a year ago

What if the time series consists entirely of 'NAs'? Is the some way to place the text in the middle of each of the boxes?

Sam Dickson · Accepted Answer · 2016-01-08 20:51:18Z

2

One option is to add the count to the variable used in the faceting:

df3$NAs <- ave(df$value,df$name,FUN=function(x) sum(is.na(x))))
df3$name1 <- paste0(df3$name,' (NA = ',df3$NAs,')') 
ggplot(data=df3, aes(time, value)) + 
  geom_line() + 
  geom_point() + facet_wrap(~ name1, nrow=1)

answered Jan 8, 2016 at 20:51

Sam Dickson

5,2892 gold badges31 silver badges48 bronze badges

1 Comment

Roland Over a year ago

Your solution is very elegant and nice, but unfortunately, it doesn't work for facet_grids, where I'd like to use this information as well.

Collectives™ on Stack Overflow

Include number of missing values in ggplot

2 Answers 2

Updated

3 Comments

1 Comment

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Updated

3 Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Linked

Related