Plotting 3 Variables on One Chart - ggplot

Question

I have some experience with base R but am trying to learn tidyverse and ggplot. I have a dataframe with 4 columns of data. I want a simple x-y plot, where the first column of data is on the x-axis, and the data in the other 3 columns is plotted on the y-axis, resulting in 3 lines on one plot. The first 15 lines of my data look like this (sorry about the image - I don't know how to insert a sample of my data):

screen shot - first 15 rows of data

I tried to plot the second and third columns of data as follows: ,

ggplot(data=SWRC_SL, aes(x=SWRC_SL$pressure_head, y=SWRC_SL$UNSODA_theta)) + 
geom_line(colour="red") + scale_x_log10() +
ggplot(data=SWRC_SL, aes(x=SWRC_SL$pressure_head, y=SWRC_SL$Vrugt_theta)) + 
geom_line(colour="blue") + scale_x_log10()

I get this error:

Error: Don't know how to add ggplot(data = SWRC_SL, aes(x = SWRC_SL$pressure_head, y = SWRC_SL$Vrugt_theta)) to a plot

I believe I should be using something like "group=" to indicate which columns should be plotted, but I haven't been able to find an example that shows how you can use gglot to plot data across multiple columns. What am I missing ?

To provide a good reproducible example, you can check this post: https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example — dc37, Jan 22 '20 at 05:32

score 2 · Answer 1 · answered Jan 22 '20 at 05:03

ggplot() is only ever called once when you create a chart. Try with the following:

ggplot() + 
  geom_line(data=SWRC_SL, aes(x=pressure_head, y=UNSODA_theta), colour="red") + 
  geom_line(data=SWRC_SL, aes(x=pressure_head, y=Vrugt_theta), colour="blue") + 
  scale_x_log10()

A better method would be to turn your data to long, where the UNSODA_theta and Vrugt_theta data are in the same column (say thetas), and have another column (say type_theta) indicating whether the data is for UNSODA_theta or Vrugt_theta. Then you could do the following:

ggplot(data=SWRC_SL, aes(x=pressure_head, y=thetas, colour=type_theta)) + 
  geom_line() + 
  scale_x_log10()

This is more desirable because ggplot2 will include a legend indicating what type of theta the colours are applied to.

Converting the data to long is the best way to go, it's what ggplot is designed to expect and it makes a lot of ggplot's features work automatically. — Marius, Jan 22 '20 at 05:14
Thanks for all the comments. I used the code provided by dc37 and the results are great. — cookie monster, Jan 23 '20 at 04:09

dc37 · Answer 2 · 2020-01-22T05:44:57.673

As suggested by @Marius, the most efficient way to plot your data is to convert them into a long format.

Using tidyverse, you can have the use of pivot_longer function (from tidyr package) and write the following code:

library(tidyverse)
SWRC_SL %>% pivot_longer(.,-pressure_head, names_to = "variable", values_to = "value") %>%
  ggplot(aes(x = pressure_head, y = value, color = variable))+
  geom_line()+
  scale_x_log10()

EDIT: Illustrating example

Using this dummy dataset:

  pressure UNSODA_theta Vrugt_theta Cassel_theta
1        0   -1.4672500   1.4119747   -2.0553118
2        1    0.5210227   0.6189239    1.4817574
3        2   -0.1587546   1.4094018    2.2796175
4        3    1.4645873   2.6888733   -0.4631109
5        4   -0.7660820   2.5865884   -1.8799346
6        5   -0.4302118   0.6690922    0.9633620

First, you pivot your data into a long format:

df %>% pivot_longer(.,-pressure, names_to = "variable", values_to = "value")

# A tibble: 45 x 3
   pressure variable      value
      <int> <chr>         <dbl>
 1        0 UNSODA_theta -1.47 
 2        0 Vrugt_theta   1.41 
 3        0 Cassel_theta -2.06 
 4        1 UNSODA_theta  0.521
 5        1 Vrugt_theta   0.619
 6        1 Cassel_theta  1.48 
 7        2 UNSODA_theta -0.159
 8        2 Vrugt_theta   1.41 
 9        2 Cassel_theta  2.28 
10        3 UNSODA_theta  1.46 
# … with 35 more rows

Now, your data are suitable for the plotting with ggplot2, you can directly add ggplot command to the previous command by adding a "pipe" (%>%) between them:

library(tidyverse)
df %>% pivot_longer(.,-pressure, names_to = "variable", values_to = "value") %>%
  ggplot(aes(x = pressure, y = value, color = variable))+
  geom_line()+
  scale_x_log10()

And you get the following plot with legend included:

Data example

structure(list(pressure = 0:14, UNSODA_theta = c(-1.46725002909224, 
0.521022742648139, -0.158754604716016, 1.4645873119698, -0.766081999604665, 
-0.430211753928547, -0.926109497377437, -0.17710396143654, 0.402011779486338, 
-0.731748173119606, 0.830373167981674, -1.20808278630446, -1.04798441280774, 
1.44115770684428, -1.01584746530465), Vrugt_theta = c(1.41197471231751, 
0.61892394889108, 1.40940183965093, 2.68887328620405, 2.58658843344197, 
0.669092199317234, -1.28523553529247, 3.49766158983416, 1.66706616676549, 
1.5413273359637, 0.986600476854091, 1.51010842295293, 0.835624168230333, 
1.42069464325451, 0.599753256022356), Cassel_theta = c(-2.05531181632119, 
1.48175740118232, 2.27961753824932, -0.46311085383842, -1.87993463341154, 
0.963361958516736, -0.0670637053409687, -2.59982761023726, 0.00319778952040447, 
-0.945450500892219, -0.511452869790608, -1.73485854395378, 2.7047128618762, 
-0.496698054586832, -2.40827011837962)), class = "data.frame", row.names = c(NA, 
-15L))

Plotting 3 Variables on One Chart - ggplot

2 Answers2