ggplot2: error when superposing two layers

Question

The following describes in a simplified manner an issue into which I run often but never get it to work. It has to do with superposing layers with different information.

Any help is appreciated.

library(dplyr)
library(ggplot2)

df = data.frame(id = c(1,1,1,1,2,2,2,3,3,3), 
            year = c(2000,2001,2002,2003,2000,2001,2002,2001,2002,2003),
            meas = c(1.1,1.2,NA,1.4,2.1,1.9,1.8,3.1,3.0,3.3))

df$id   = as.factor(df$id)
df$year = as.factor(df$year)

tib = as.tbl(df)

which produces the tibble:

# A tibble: 10 x 3
       id   year  meas
   <fctr> <fctr> <dbl>
 1      1   2000   1.1
 2      1   2001   1.2
 3      1   2002    NA
 4      1   2003   1.4
 5      2   2000   2.1
 6      2   2001   1.9
 7      2   2002   1.8
 8      3   2001   3.1
 9      3   2002   3.0
10      3   2003   3.3

Presently, I compute the median of the meas-ures for each year in the cross section.

tib2 = tib %>% group_by(year) %>%       
    summarise(medi=median(meas,na.rm=T))
tib2

Which gives as expected

# A tibble: 4 x 2
    year  medi
  <fctr> <dbl>
1   2000  1.60
2   2001  1.90
3   2002  2.40
4   2003  2.35

So far so good. Now, I create my first plot

p = ggplot(tib, aes(x = year, y = meas, group = id, color = id)) p = p + geom_line() p which produces

On the other hand, the following produces a set of fat dots as it should

p1=ggplot(tib2, aes(x = year, y = medi)) + geom_point(colour = 'red', size = 3)
p1

I try however to place the red dots in the first figure like so:

p = p + geom_point(data = tib2, aes(x = year, y = medi), colour = 'red', size = 3)
p

This is where I get the error:

Don't know how to automatically pick scale for object of type function. Defaulting to continuous. Error in (function (..., row.names = NULL, check.rows = FALSE, check.names = TRUE, the arguments imply different numbers of lines : 4, 0

Remark: I noticed that the missing value induces a cut in the red line which I might wish to interpolate. When I replace the NA by some value, the error persists however.

shouldn't that last code chunk be p = p + geom_point(...)? — user3603486
– user3603486, Commented May 8, 2018 at 9:53
The "object of type function" error suggests that aes(x = year...) might be misinterpreted (is lubridate loaded by any chance?) What happens if you do ggplot() + geom_line(data = tib2, aes(x = year, y = medi), colour = 'red', size = 3) — user3603486
– user3603486, Commented May 8, 2018 at 9:55
Lubridate is not on. I made a mistake copying. The error persists with p=p+geom_point. — Michael Rockinger
– Michael Rockinger, Commented May 8, 2018 at 10:00
ggplot()+geom_point(data = tib2, aes(x = year, y = medi), colour = 'red', size = 3) produces the second figure. It is when I add p=p+geom_point(data = tib2, aes(x = year, y = medi), colour = 'red', size = 3) that I get the bug. Thanks — Michael Rockinger
– Michael Rockinger, Commented May 8, 2018 at 10:03

micstr · Accepted Answer · 2023-03-07 11:04:00Z

3

The "object of type function" error relates to the dplyr::id() function.

When you assign an aesthetic in the ggplot call it applies to all the plot, and while it is correctly interpreted in the first geom in the tib dataframe context, the tib2 context does not have it, so it produces an error.

You can pass the relevant aes to the single geom, or you can use inherit.aes = FALSE on the 'tib2 geom':

library(ggplot)

ggplot(tib) + 
  geom_line(aes(x = year, y = meas, group = id, color = id)) +
  geom_point(data = tib2, aes(x = year, y = medi), size = 3, color = 'red')


ggplot(tib, aes(x = year, y = meas, group = id, color = id)) + 
  geom_line() +
  geom_point(data = tib2, aes(x = year, y = medi), size = 3, color = 'red', inherit.aes = FALSE)

Or even more elegantly, specify the aes that are common in the ggplot call and the others where needed:

ggplot(tib, aes(x = year)) + 
  geom_line(aes(y = meas, group = id, color = id)) +
  geom_point(data = tib2, aes(y=medi), size = 3, color = 'red')

Created on 2018-05-08 by the reprex package (v0.2.0).

edited Mar 7, 2023 at 11:04

micstr

5,2568 gold badges52 silver badges80 bronze badges

answered May 8, 2018 at 10:31

GGamba

13.7k3 gold badges41 silver badges51 bronze badges

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

2 Comments

Michael Rockinger Over a year ago

To avoid mistakes, would it make sense to follow the following rule: put the first dataframe into ggplot. Put aesthetics in geom as they come along. To change dataframe, put new dataframe+aes in new geom?

GGamba Over a year ago

That's what I usually do. Also changing data in geom should be done sparingly

Roman · Accepted Answer · 2018-05-08 09:59:04Z

2

you can try

ggplot(tib, aes(x = year, y = meas, group = id, color = id)) +
   geom_line() +
   stat_summary(fun.y = median, geom="point", aes(group=year), colour = 'red', size = 3)

or use as Dash2 recommended

ggplot() +
   geom_line(data=tib, aes(x = year, y = meas, group = id, color = id)) +
   geom_point(data = tib2, aes(x = year, y = medi), colour = 'red', size = 3)

answered May 8, 2018 at 9:59

Roman

17.7k3 gold badges39 silver badges52 bronze badges

1 Comment

Michael Rockinger Over a year ago

thanks! This fixes the bug. The mistake must be that by defining within ggplot the data = tib it persists as I introduce the new data. This implies one cannot just add layers, from the beginning the intention should be clear. Unless there is a parameter I missed. Thanks again.

Collectives™ on Stack Overflow

ggplot2: error when superposing two layers

2 Answers 2

2 Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related