CODEWITHSUNDEEP
rggplot2
Tyler Wood
Tyler Wood

Reputation: 1967

ggplot within function can't find dataframe passed to function

I have the following function that I'd like to use to generate density histograms of various features in a dataset:

fraudSplitHist <- function(dataframe, stat_col, signal_col='fraud') {
  hist_data <- dataframe[,c(stat_col, signal_col)]
  colnames(hist_data) <- c('stat', 'fraud')

  hist_data$stat <- winsor1(hist_data$stat) # remove outliers

  ggplot(hist_data, aes(hist_data$stat, fill=as.factor(hist_data$fraud))) +
    geom_histogram(alpha=0.5, aes(y=..density..), position='identity', binwidth=1)
}

Running it always give me this error 

fraudSplitHist(dataframe=data, stat_col='some_column_of_values')
Error in eval(expr, envir, enclos) : object 'hist_data' not found

However, it works with qplot

fraudSplitHist <- function(dataframe, stat_col, fraud_col='fraud') {
  hist_data <- dataframe[,c(stat_col, fraud_col)]
  colnames(hist_data) <- c('stat', 'fraud')

  hist_data$stat <- winsor1(hist_data$stat)

  qplot(data=hist_data, x=stat, geom="histogram", fill=as.factor(fraud), binwidth=10,
        position='identity')
}

Any ideas what I could be missing here?

Upvotes: 2

Views: 704

Answers (1)

MrFlick
MrFlick

Reputation: 206207

You should not specify the data.frame name in your aes() settings. You should be just using column names from the data.frame. For example, this should work

simpletest <- function(dataframe, x_col, y_col='cyl') {
  plot_data <- dataframe[,c(x_col, y_col)]
  colnames(plot_data) <- c('stat', 'fraud')

   ggplot(plot_data, aes(stat, fill=as.factor(fraud))) +
    geom_histogram(alpha=0.5, aes(y=..density..), position='identity', binwidth=1)
}

simpletest(mtcars, "disp")

Upvotes: 3

Related Questions