How to use quosure in ggplot to generate plotting function

Question

I have a data set where I need to generate multiple plots based on different columns, so instead of copy/pasting, I am trying to write a function, where I pass the function the dataframe and the column name. I am pretty sure that this requires a quosure, but I don't seem to be able to get it to work in ggplot2.

Ideally, I would be able to pass the dataframe, RawData, and the column name (B in this case) and it would generate a bar plot using column B, and including the column B label on the plot. For example.

RawData = data.frame(A = runif(12, 0, 5), B = LETTERS[seq(1,4)], C = letters[seq(3,1)])

ggplot(RawData, aes(x=A,y=B))+
  geom_bar(stat="identity",position="dodge")+
  xlab("Group")+ylab("B")

The function I wrote is effectively this...

PlotFn  = function(df, VarName) {
  VarName = quo(VarName)
  ggplot(df, aes(x=A,y=!!VarName))+
  geom_bar(stat="identity",position="dodge")+
  xlab("Group")+ylab(as_label(!!VarName))
}

PlotFn(RawData, B)

Which throws the error:

Error: Quosures can only be unquoted within a quasiquotation context.

# Bad: list(!!myquosure)

# Good: dplyr::mutate(data, !!myquosure)

Ideally I would be able to have my plot function run

PlotFn(RawData, B)

PlotFn(RawData, C)

and give bar plots that are labeled appropriately, like the one here.

Good review here: tidyverse.org/articles/2018/07/ggplot2-tidy-evaluation — David Klotz
– David Klotz, Commented Aug 6, 2019 at 17:18

Tung · Accepted Answer · 2019-08-06 17:14:31Z

3

You don't need to unquote (!!) VarName

library(rlang)
library(tidyverse)

RawData = data.frame(A = runif(12, 0, 5), 
                     B = LETTERS[seq(1,4)], 
                     C = letters[seq(3,1)])

PlotFn <- function(df, .VarName) {
  VarName <- enquo(.VarName)
  ggplot(df, aes(x = A, y = !!VarName)) +
    geom_bar(stat = "identity", position = "dodge") +
    ylab(as_label(VarName)) +
    xlab("Group")
}
PlotFn(RawData, C)

I personally prefer .data[[]] like this

PlotFn2 <- function(df, .VarName) {
  ggplot(df, aes(x = A, y = .data[[.VarName]])) +
    geom_col(position = "dodge") +
    ylab(.VarName) +
    xlab("Group")
}

plot_list <- c("B", "C") %>% 
  map(~ PlotFn2(RawData, .x))
plot_list
#> [[1]]

#> 
#> [[2]]

^{Created on 2019-08-06 by the reprex package (v0.3.0)}

answered Aug 6, 2019 at 17:14

Tung

28.6k8 gold badges98 silver badges129 bronze badges

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

4 Comments

TimTeaFan Over a year ago

Another option you could add to your answer is PlotFn <- function(df, .VarName) { ggplot(df, aes(x = A, y = {{ .VarName. }})) }.

Tung Over a year ago

@TimTeaFan: thanks for the reminder! I kept forgetting about the new rlang's function

EBrewe Over a year ago

@Tung, thank you, this is pretty confusing, somehow I couldn't get this to go the way I thought it might. You use the rlang to say .VarName right? I just tried to run with

PlotFn  = function(df, VarName) { VarName = enquo(VarName) ggplot(df, aes(x=A,y=!!VarName))+   geom_bar(stat="identity",position="dodge")+    xlab("Group")+ylab(as_label(VarName))  }

and it worked, but didn't need rlang, am I right on this?

Tung Over a year ago

@EBrewe: no you don't need rlang. Only needed if you use curly-curly {{ }} as TimTeaFan mentioned above

Collectives™ on Stack Overflow

How to use quosure in ggplot to generate plotting function

1 Answer 1

4 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

4 Comments

Your Answer

Sign up or log in

Post as a guest

Related