Writing a function for ggplot

Question

I am new to R and have been trying to figure this out for a while. Basically, I have a data frame, and various y variables. I am trying to write a function that will allow me to come up with a customized graph template for the many different y variables that I have. I am trying the following code below but I am met with this error:

1: In eval(expr, envir, enclos) : NAs introduced by coercion

2: In aes_string(xvar[max(which(complete.cases(yvar)))], yvar[max(which(complete.cases(yvar)))], : NAs introduced by coercion

The code works if I add the variables in directly and not through a function. I believe that it is something to do with how the function plugs in the xvar into the as.numeric() function. I am not sure but any of you knows how to deal with this?

test <- function (Data, xvar, yvar){
  # Plot data
  plot <- ggplot(subset(Data,!is.na((yvar))), aes_string(xvar, yvar)) + geom_line(colour="darkblue") + theme_bw()
  # Add Trendline for recent data
  plot <- plot + geom_smooth(data=subset(Data, xvar > as.numeric(xvar)[max(which(complete.cases(yvar)))-8]), method = "lm")
  # Label most recent data
  plot + geom_text(data = Data, aes_string(xvar[max(which(complete.cases(yvar)))],
                             yvar[max(which(complete.cases(yvar)))],
                             label = as.numeric(yvar)[max(which(complete.cases(yvar)))],
                             hjust= -0.5, vjust = 0.5))

aes_string expects a character string. xvar[max(which(complete.cases(yvar))) doesn't make sense there. Compute this outside of ggplot2. Also, do not use subset inside functions. It's documentation warns against this. — Roland
– Roland, Commented May 31, 2016 at 9:05

lukeA · Accepted Answer · 2016-05-31 09:08:32Z

1

As xvar is probably (you do not show a reproducible example) a character vector of length 1, subsetting like xvar[] will not yield the desired result.

You could try something like

library(ggplot2)
f <- function(data, xvar, yvar) {
  ggplot(data, aes_string(xvar, yvar)) + 
    geom_point() + 
    geom_smooth(data=subset(data, eval(parse(text=xvar)) > 5), method = "lm")
}

or

f <- function(data, xvar, yvar) {
  ggplot(data, aes_string(xvar, yvar)) + 
    geom_point() + 
    geom_smooth(data = data[data[, xvar]>5, ], method = "lm")
}

f(mtcars, "cyl", "disp")

answered May 31, 2016 at 9:08

lukeA

54.4k5 gold badges102 silver badges101 bronze badges

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

1 Comment

Binggg Over a year ago

I still don't get how it can be corrected? I understand that it's a character vector of length 1 (is this the same as a character string?). Assuming I have a data of such initially: > X1 <- as.yearqtr(seq(as.Date("2010/3/1"), by = "quarter", length.out = 10)) > Y1 <- as.vector(c(124,315,363,574,345,434,141,512,142,647)) > Data1 <- data.frame(X1, Y1)

Pete Barwis · Accepted Answer · 2016-06-03 01:54:40Z

I think @LukeA has gotten you practically all the way there, but here is an example that uses your data and adds a few more columns to help demonstrate how you can pass column names into ggplot inside your own function.

It uses your variable names. It subsets your data into a data.frame with non-missing values for y, and then it subsets your data into a separate data.frame that allows you to add additional filtering criteria to your smoothing function.

library(zoo)
set.seed(72)

X1 <- as.yearqtr(seq(as.Date("2010/3/1"), by = "quarter", length.out = 10))
Y1 <- as.vector(c(124,315,363,574,345,434,141,512,142,647))
Y2 <- sample(Y1)
Y3 <- sample(Y1)

Data1 <- data.frame(X1, Y1, Y2, Y3) 


plot_function <- function(data, xvar, yvar){

  # remove rows with NA on yvar
  mydata1 <- data[!is.na(data[, yvar]), ]

  # remove rows with NA on yvar and subset yvar above some threshold
  mydata2 <- data[!is.na(data[, yvar]) & data[, yvar] > 400, ]

  # plot it
  myplot <- ggplot(mydata1, aes_string(xvar, yvar)) + 
    geom_line(colour="darkblue") + 
    scale_x_yearqtr(limits = c(min(mydata1[, xvar]), max(mydata1[, xvar])), format = "%YQ%q") +
    geom_smooth(data = mydata2, aes_string(xvar, yvar), method = "lm") +
    geom_text(data = mydata1, aes_string(xvar, yvar, label = yvar),  hjust= -0.5, vjust = 0.5) +
    theme_bw()

  return(myplot)
}


plot_function(data = Data1, xvar = "X1", yvar = "Y1")
plot_function(data = Data1, xvar = "X1", yvar = "Y2")
plot_function(data = Data1, xvar = "X1", yvar = "Y3")

Collectives™ on Stack Overflow

Writing a function for ggplot

2 Answers 2

1 Comment

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Related