String as column name in ggplot

[esa_chris]

I want to use a string instead of a column name in a plot.

library("ggplot2")

df <- mtcars
df$rown <- seq.int(nrow(df))

variable = mpg

p <- ggplot() +  
  geom_line(data = df, aes(x=rown, y=mpg)) 

p

works just fine. However, I wish to replace 'mpg' by 'variable' such that I can simply change the latter (e.g. to hp).

[AyushBipinPatel]

Hey @esa_chris ,

you are almost there. You can try the following:

library("ggplot2")

df <- mtcars
df$rown <- seq.int(nrow(df))

variable = "mpg"  # needs to be a string, which can then be changed.
ggplot() +  
  geom_line(data = df, aes(x=rown, y=mpg)) +
  labs(
    y = variable, # add title for y axis,
    title = "PLot title", # add plot title
    )

There are other arguments in the labs() function as well, such as subtitle of the plot, caption and x axis title. Above is one way of doing it. However, if you want to avoid assigning a value to a named object, such as variable in this case, you can try the following.

library("ggplot2")

df <- mtcars
df$rown <- seq.int(nrow(df))

create_chart <- function(data,x_var,y_var){
  
  data|>
  ggplot() +  
    geom_line(aes(x= .data[[x_var]], y= .data[[y_var]])) +
    labs(
      y = y_var, # add title for y axis,
      title = "PLot title", # add plot title
    )
  
}

create_chart(data = df,x_var = "rown",y_var = "mpg")
create_chart(data = df,x_var = "rown",y_var = "hp")

Hope this helps,
Ayush

[esa_chris]

Thanks for the swift response. I suppose I did not make myself very clear. I do not want to add captions, axis titles etc. What I had in mind is something like this

library("ggplot2")

df <- mtcars
df$rown <- seq.int(nrow(df))

variable = mpg

p <- ggplot() +  
  geom_line(data = df, aes(x=rown, y=variable)) 

p

But that does not work

[AyushBipinPatel]

Hi @esa_chris , I miss read your question, sorry for that. I have edited the previous answer. It has two approaches you can use.

Best,
Aysuh

[esa_chris]

I appreciate your feedback, but this is not really what I need. The variable I want to plot on the y-axis changes dynamically. It is not fruitful to manually change it as the name of the variable is the result of previous steps.

Is there no way to use a string that holds the name of the column that I want to use on the y-axis?

variable = mpg

p <- ggplot() +  
  geom_line(data = df, aes(x=rown, y=mpg)) 

instead of 'mpg' in the last row I want to use a string. The value of the string will be assigned in a previous steps.

variable = mpg

p <- ggplot() +  
  geom_line(data = df, aes(x=rown, y=variable)) 

Like I simply want to use the "value" of the string "variable" in the above example

[AyushBipinPatel]

Hi @esa_chris ,

@esa_chris

instead of 'mpg' in the last row I want to use a string. The value of the string will be assigned in a previous steps.

The above can be done by using the function shown in the previous reply.

library("ggplot2")

df <- mtcars
df$rown <- seq.int(nrow(df))

create_chart <- function(data,x_var,y_var){
  
  data|>
  ggplot() +  
    geom_line(aes(x= .data[[x_var]], y= .data[[y_var]])) +
    labs(
      y = y_var, # add title for y axis,
      title = "PLot title", # add plot title
    )
  
}

variable <- "mpg" # a string that holds the name of the varible

create_chart(data = df,x_var = "rown",y_var = variable) # plots for rown vs mpg

variable <-  "hp" # change the column name to be plotted on y axis

create_chart(data = df,x_var = "rown",y_var = variable) # plots for rown vs hp

Hope this is helps,
Ayush

[DavoWW]

Hi @esa_chris,
As @AyushBipinPatel has demonstrated, you can use the {rlang} .data pronoun to set the column name via variable with no need for a user-defined function:

library(ggplot2)
df <- mtcars
df$rown <- seq.int(nrow(df))

variable <- "mpg"  # Beware, mpg is a separate inbuilt ggplot2 dataset

p <- ggplot() +  
  geom_line(data = df, aes(x=rown, y=.data[[variable]])) 
p

Happy Holidays!

[esa_chris]

Both approaches work, thanks

[esa_chris]

I have a follow up question.

How can a specify the actual color of the different lines manually using the function approach? The usual scale_color_manual does not work as it dose not accept the arguments of the function as an input. There is not error, but all the lines are in shades of grey.