RStudio provides a nice function View
(with uppercase V
) to take a look into the data, but with R it's still nasty to get orientation in a large data set. The most common options are...
names(df)
str(df)
If you're coming from SPSS, R seems like a downgrade in this respect. I wondered whether there is a more user-friendly option? I did not find a ready-one, so I'd like to share my solution with you.
Using RStudio's built-in function View
, it's white simple to have a variable listing for a data.frame
similar to the one in SPSS. This function creates a new data.frame
with the variable information and displays in the RStudio GUI via View
.
# Better variables view
Varlist = function(sia) {
# Init varlist output
varlist = data.frame(row.names = names(sia))
varlist[["comment"]] = NA
varlist[["type"]] = NA
varlist[["values"]] = NA
varlist[["NAs"]] = NA
# Fill with meta information
for (var in names(sia)) {
if (!is.null(comment(sia[[var]]))) {
varlist[[var, "comment"]] = comment(sia[[var]])
}
varlist[[var, "NAs"]] = sum(is.na(sia[[var]]))
if (is.factor(sia[[var]])) {
varlist[[var, "type"]] = "factor"
varlist[[var, "values"]] = paste(levels(sia[[var]]), collapse=", ")
} else if (is.character(sia[[var]])) {
varlist[[var, "type"]] = "character"
} else if (is.logical(sia[[var]])) {
varlist[[var, "type"]] = "logical"
n = sum(!is.na(sia[[var]]))
if (n > 0) {
varlist[[var, "values"]] = paste(round(sum(sia[[var]], na.rm=T) / n * 100), "% TRUE", sep="")
}
} else if (is.numeric(sia[[var]])) {
varlist[[var, "type"]] = typeof(sia[[var]])
n = sum(!is.na(sia[[var]]))
if (n > 0) {
varlist[[var, "values"]] = paste(min(sia[[var]], na.rm=T), "...", max(sia[[var]], na.rm=T))
}
} else {
varlist[[var, "type"]] = typeof(sia[[var]])
}
}
View(varlist)
}
My recommendation is to store that as a file (e.g., Varlist.R
) and whever you need it, just type:
source("Varlist.R")
Varlist(df)
Again please take note of the uppercase V
using as function name.
Limitation: When working with data.frame
, the listing will not be updated unless Varlist(df)
is run again.
Note: R has a built-in option to view data with print
. If working with pure R, just replace the View(varlist)
by print(varlist)
. Yet, depending on screen size, Hmisc::describe()
could be a better option for the console.