Geom_line missing the path?

tidyverse
1

geom_line doesn't follow the same path as geom_text with the same data
it goes as if its data were d instead of d2
Does anybody would help me to see what I'm doing wrong? please
'''
d=data.frame(height=c(1,2,2,3,4), weight=c(1,3,4,4,3))
d2=data.frame(h=c(1,2,3,2,4), w=c(1,3,4,4,3))
ggplot() +
geom_point(data=d, mapping = aes(x=height, y=weight), size=8,shape=21, fill='white') +
geom_text(data=d2, mapping = aes(x=h, y=w), label=seq(1, nrow(d2))) +
geom_line(data=d2, mapping = aes(x=h, y=w))

'''

2

Rplot
Rplot709×573 7.98 KB

the line doesn't follow the numbers

3

Is this explanation correct?

geom_line reorders data by x before representing data as if it is a function
so y=f(x) is a function because there is a sigle y for each x

4

geom_line connects points in order from the lowest the x value to the highest. If an x value is repeated, the repeated values are kept (relative to each other) in the order they appear in the data.

To connect all points in the order they appear in the data, use geom_path.

geom_text isn't following a path. It's plotting a label at each (h, w) location in the data. If the label comes from the data, that is, a data column is mapped to the label aesthetic, which would be the typical case, then the label always corresponds to the appropriate (h,w) value. However, if the label values are hardcoded outside of aes, as in your example, then the labels will be plotted in the hardcoded order, which, in general, will be different than the sorted order of the x values, unless the data frame is already sorted by the x values before plotting. The advantage of using aesthetic mappings is that they intrinsically maintain the appropriate correspondences between all the data values.

Below are a few examples where we include both geom_point and geom_path.

library(tidyverse)
theme_set(theme_classic())

d=data.frame(height=c(1,2,2,3,4), weight=c(1,3,4,4,3))
d2=data_frame(h=c(1,2,3,2,4), w=c(1,3,4,4,3),
              labs=paste0(h,c("","a","","b","")))

First, we'll plot the data with the labs column mapped to the label aesthetic.

ggplot(d2, aes(h, w)) +
  geom_point(data=d, aes(x=height, y=weight), size=8, shape=21, fill='white') +
  geom_line() +
  geom_path(colour="red", linetype=2, size=1.5) +
  geom_text(aes(label=labs), colour="blue")

Rplot18

Now let's hardcode the label values outside of aes:

ggplot(d2, aes(h, w)) +
  geom_point(data=d, aes(x=height, y=weight), size=8, shape=21, fill='white') +
  geom_line() +
  geom_path(colour="red", linetype=2, size=1.5) +
  geom_text(colour="blue", label=c(1,"2b",3,"2a",5))

Rplot20

Finally, let's modify the data frame by switching the order of the points with height=2, without changing the order of any other rows:

d2 = d2[c(1,4,3,2,5), ]

ggplot(d2, aes(h, w)) +
  geom_point(data=d, aes(x=height, y=weight), size=8, shape=21, fill='white') +
  geom_line() +
  geom_path(colour="red", linetype=2, size=1.5) +
  geom_text(aes(label=labs), colour="blue")

Rplot19

3 Likes
5

joels, thank you very much for your explanation - I'm just beginning to learn how to use this package