Label the points as date in the proper format in ggplot in R

Question

I am plotting a line chart using ggplot2 in R. I want to name the points above a certain threshold in the proper date format.

My code for plotting the graph is:

ggplot(DateSubset1, aes(TimeStamp)) + 
    geom_line(aes(y = CPU, colour = "Orange")) + 
    geom_line(aes(y = MEM), colour = "Black")+
    scale_x_datetime(date_break = "1 days")+
    geom_point(aes (x= TimeStamp, y=CPU), size = 1,colour = "Purple", 
        subset(DateSubset1, CPU>25 ))+
    geom_point(aes (x= TimeStamp, y=MEM), size = 1,colour = "Blue", 
        subset(DateSubset1, MEM>10 ))+
    scale_y_continuous(breaks = c(5,10,15,20,25,30,35,40,45,50,55,60,65,70,75,80))

My Graph looks like this:

I want to label these points (which are above a certain threshold) with proper date format as my dataset has. I have tried

geom_text(aes(y=CPU, label= ifelse(CPU>25, TimeStamp, '')))

Using this my graph looks like:

And

geom_text(aes(y= CPU,label= ifelse(CPU>25, format(TimeStamp), format = 
"%y%m%d %h%m%s",'')))

And

geom_text(aes(y= CPU, label=ifelse(CPU>25, as.Date(TimeStamp), '')))

And

geom_text(aes(y= CPU, label=ifelse(CPU>25, as.Date.POSIXct(TimeStamp), '')))

String of dataset:

data.frame':    
1420 obs. of  3 variables:
$ TimeStamp: POSIXct, format: "2017-06-28 07:03:02" "2017-06-28 07:06:01" 
"2017-06-28 07:09:01" ...
$ CPU      : num  0.9 0.8 12.2 3.7 2.3 1.7 1.4 1.1 1 0.9 ...
$ MEM      : num  1.7 1.8 1.5 1.8 1.8 1.8 1.9 1.9 1.9 2.1 ...

The sample Data looks like:

TimeStamp                CPU   MEM
2017-06-28 07:03:02      0.9   1.7
2017-06-28 07:06:01      0.8   1.8
2017-06-28 07:09:01      12.2  1.5
2017-06-28 07:12:01      3.7   1.8
2017-06-28 07:15:01      2.3   1.8

Please provide example data so that your figure can be generated by others. — B Williams, Jul 25 '17 at 23:31
Possible duplicate of [Label points in geom\_point](https://stackoverflow.com/questions/15624656/label-points-in-geom-point) — Paul Endymion, Jul 25 '17 at 23:42
Not a duplicate because the issue is formatting the text label as a date — user101089, Jul 26 '17 at 01:33
Example data, please! What does 'TimeStamp' look like? Are you able to convert timestamp to a POSIXct outside the `geom_text` call? — Remko Duursma, Jul 26 '17 at 02:23
the string of dataset is data.frame': 1420 obs. of 3 variables: $ TimeStamp: POSIXct, format: "2017-06-28 07:03:02" "2017-06-28 07:06:01" "2017-06-28 07:09:01" ... $ CPU : num 0.9 0.8 12.2 3.7 2.3 1.7 1.4 1.1 1 0.9 ... $ MEM : num 1.7 1.8 1.5 1.8 1.8 1.8 1.9 1.9 1.9 2.1 ... — Sapan Badjatiya, Jul 26 '17 at 04:07

AK88 · Accepted Answer · 2017-07-26T06:25:00.317

0

OK, try this code:

zz = '
   CPU   MEM
   0.9   1.7
   0.8   1.8
   12.2  1.5
'

df <- read.table(text = zz, header = TRUE)
df

TmS = c("2017-06-28 07:03:02", "2017-06-28 07:06:01", "2017-06-28 07:09:01")
df = cbind(TmS, df)
df$TmS = as.character(df$TmS)

label = as.character(ifelse(df$CPU>10, df$TmS, ''))
df$TmS = as.POSIXct(df$TmS)


ggplot(df, aes(TmS)) + 
  geom_line(aes(y = CPU, colour = "Orange")) + 
  geom_line(aes(y = MEM), colour = "Black")+
  scale_x_datetime(date_break = "1 days")+
  geom_point(aes (x= TmS, y=CPU), size = 1,colour = "Purple", 
             subset(df, CPU>10))+
  geom_point(aes (x= TmS, y=MEM), size = 1,colour = "Blue", 
             subset(df, MEM>1.5))+
  scale_y_continuous(breaks = c(5,10,15,20,25,30,35,40,45,50,55,60,65,70,75,80))+
  geom_text(aes(y= CPU, label=label))

edited Jul 26 '17 at 06:25

answered Jul 26 '17 at 02:47

AK88

2,946
2
12
31

When I tried this I am seeing this error- Error in unclass(x)/86400 : non-numeric argument to binary operator – Sapan Badjatiya Jul 26 '17 at 04:28
I made changes. Let me know if you have further questions. – AK88 Jul 26 '17 at 06:25
Yeah I got it. Thanks!! Is there any way in which I can avoid the overlapping of labels? – Sapan Badjatiya Jul 26 '17 at 18:30
Well, I don't think there is an easy solution. Have a look here: https://stackoverflow.com/questions/21585128/how-to-avoid-overlapping-labels-with-identical-data-points-in-scatterplot-ggpl – AK88 Jul 27 '17 at 03:03
I got it. We can use ggrepel package. and using geom_text_repel in place of geom_text – Sapan Badjatiya Jul 27 '17 at 23:53
We can also use plotly package, which makes the graph pretty interactive – Sapan Badjatiya Jul 27 '17 at 23:55

Label the points as date in the proper format in ggplot in R

1 Answers1