While analyzing data sets, I sometimes stop and wonder how a user-defined function might automate a task that I see myself carrying out frequently.  An example came up recently that involved something that I used to do with every new data set back when I used SPSS (I think it was the “Explore” function in SPSS…): visualize all the categorical data using barplots and the continuous data using histograms. 

To that end, I started to wonder if I could use for to loop over an index within a dataframe to carry this out.  So I gave writing the loop a shot and failed miserably; however, I think I had enough of a well-formed attempt to post on stack overflow (http://stackoverflow.com/questions/10643841/simple-barplot-of-variables-in-df-using-a-loop). 

My attempt looked like this:

y <- LETTERS[as.integer(rnorm(100, mean=5, sd=1))]

z <- LETTERS[as.integer(rnorm(100, mean=10, sd=1))]

x <- round(rnorm(100, mean=5, sd=2.5),2) data <- as.data.frame(cbind(x,y,z))

A<- do.call(“cbind”, lapply(data, class))

B<- as.vector( A[1,])

C <- grep(“character|factor”, B)

for (i in 1:length(C)) {   x <- C[i]   counti <- table(data[,x])   y <- barplot(counti, main=paste(“Barplot for var”, x))   return(y)}

Here’s hoping that those much smarter than me can come to my rescue…