Multi tool use

R: Reshape data to create a multiline-plot from different variables

I have to create a plot with different lines from different variables.

One the y-axis:

> a1
[1] 7 10 12 19 30 45 68 104 151 214 279 362 425 467 487 493 498 500
> a2
[1] 7 11 20 33 59 94 151 244 334 425 483 500
> a3
[1] 7 17 35 77 150 241 361 450 488 500
> a4
[1] 7 20 50 129 292 441 498 500
> a5
[1] 7 22 68 191 382 493 500


On the x-axis:

> b1
[1] 18
> b2
[1] 12
> b3
[1] 12
> b4
[1] 10
> b5
[1] 8


a1 and b1 should be one line, a2 and b2 another and so on.

How can I fit the data to the same scale?

How can I generate a plot in which all lines are included?

Right now, the single plots look like this:
The code used for the plots looks like this:

plot(a1, type="l",col="red",xlim=c(1,b1),main="Example1", xlab="x", ylab = "y")
plot(a2,type="l",col="blue",xlim=c(1,b2),main="Example2", xlab="x", ylab = "y")


The requestet dput():

> dput(a1)
c(6, 8, 12, 20, 34, 54, 80, 110, 159, 214, 281, 345, 402, 447, 479, 492, 494, 498, 500)
> dput(b1)
19L
> dput(a2)
c(6, 12, 24, 42, 76, 127, 209, 306, 375, 441, 476, 495, 499, 499, 500)
> dput(b2)
15L


As you can probably tell, I am very new to using R.

dput

ggplot

b1

b2

2 Answers
2

This is really a problem of getting data in shape in order to plot it. I'm a little confused by what the data is exactly that you're working with, i.e. what b1, b2, etc. are. But with the a1 and a2 you posted (and whatever similar vectors you have), you should get everything into a data frame. The fact that these vectors are of different lengths doesn't matter because ggplot expects long-shaped data anyway. That is, as far as I can tell, a1 and a2 contain different sets of measurements of the same sort of thing (stock values, size of an organism, probability of some outcome, etc); therefore, in the tidy-data paradigm, they should be different groups of the same variable.

b1

b2

a1

a2

ggplot

a1

a2

So I first make a data frame/tibble of each set of data, giving them an x that's just integers from 1 to whatever is the length of its a vector, then give it a group marker so I can keep each set of measurements straight. Then I bind that into a single data frame.

x

a

library(tidyverse)

df1 <- tibble(
a = c(6, 8, 12, 20, 34, 54, 80, 110, 159, 214, 281, 345, 402, 447, 479, 492, 494, 498, 500),
x = 1:length(a),
group = 1
)
df2 <- tibble(
a = c(6, 12, 24, 42, 76, 127, 209, 306, 375, 441, 476, 495, 499, 499, 500),
x = 1:length(a),
group = 2
)
df <- bind_rows(df1, df2)
df
#> # A tibble: 34 x 3
#> a x group
#> <dbl> <int> <dbl>
#> 1 6 1 1
#> 2 8 2 1
#> 3 12 3 1
#> 4 20 4 1
#> 5 34 5 1
#> 6 54 6 1
#> 7 80 7 1
#> 8 110 8 1
#> 9 159 9 1
#> 10 214 10 1
#> # ... with 24 more rows


The ggplot way of doing things is to map variables onto aesthetics such as color. If you find yourself making multiple geoms that serve essentially the same purpose (like a geom_line for each of a set of observations), you probably need to reshape your data.

ggplot

geom_line

ggplot(df, aes(x = x, y = a, color = as.factor(group), group = as.factor(group))) +
geom_line()


This will scale for however many sets of observations you have—create smaller data frames for each a*, mark them as a group, and bind them all together.

a*

As with many things ggplot, there are a lot of ways to do this. I think the easiest would be to save all your vectors in one data.frame/tibble and then refer to the columns in individual geom_line calls. In the example below I made each line a different color just to more easily distinguish between the two.

ggplot

geom_line

library(tidyverse)

data <- tibble(a1 = c(1, 2 ,3, 4, 5), b1 = c(20), a2 = c(5, 6, 7, 8, 9), b2 = c(30))

ggplot(data = data) +
geom_line(aes(x = a1, y = b1), color = "blue") +
geom_line(aes(x = a2, y = b2), color = "red")


Or with different length data:

library(tidyverse)

data_1 <- tibble(a1 = c(1, 2 ,3, 4, 5), b1 = c(20))

data_2 <- tibble(a2 = c(5, 6, 7, 8, 9, 10, 11, 12), b2 = c(30))

ggplot() +
geom_line(data = data_1, aes(x = a1, y = b1), color = "blue") +
geom_line(data = data_2, aes(x = a2, y = b2), color = "red")


By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.