passing string to ggplot function

Question

I would like to have a function that i could apply to any object that meets a criteria, and have a nice ggplot scatter plot with regression line print.

How

Answer 1

You can specify the columns using indexes.

> library(ggplot2)
> df <- data.frame(a = 1:10, b = 1:10, c=c("x"))
> ggplot(data = df, aes(x = df[,1], y = df[,2])) + geom_line()

Answer 2

Your function throws an error since aes() tries to evaluate the argument in the column names of your data. To be more specific, aes() tries to evaluate colnames(tempData)[1] as a column name and this column doesn't exist.

To fix this, you somehow have to tell ggplot that you are not passing a column name, but an expression (a string) that will resolve to a column name.

Use aes_string() for this. Concretely, simply replace aes() with aes_string(). Try this:

PointReg <- function(Xts, a=1, b=2) {
  stopifnot(is.xts(Xts), 
            ncol(Xts) >1)
  tempData <- Xts[, c(a,b)]
  gPlot <- ggplot(data = tempData, 
                  aes_string(x = colnames(tempData)[1],
                      y = colnames(tempData)[2])) +
    geom_point(shape=1) +
    geom_smooth(method = lm)
  gPlot
}