R/summary_r.R

In UBC-MDS/CleanR: Understand your messy data

#' Summary
#'
#' @description
#' It will determine the main data type of the column by calculated the type
#' of each row entry in the column, and using the most frequent data type as
#' the expected input for that column. It will perform two different summary
#' statistics based on 2 different groups of datatypes of either 1) string or
#' 2) int/float. For number columns it returns a dataframe of columns which
#' contain dataframes of summary statistics including mean value for each column,
#' min, max, count (number of non NA values per column) and count_NA (number of
#' NA values per column). Similarly, for string columns it returns the unique
#' string values and their counts. It will also provide a count of NA values
#' which will include empty strings, and anything other than the correct data
#' type for each column.
#'
#' @param data (tbl_df, df, data.frame) dataframe that the function will use to locate NAs
#'
#' @return a nested list of columns that includes all the summary statistics information
#' @importFrom stats median
#' @export
summary_r <- function(data) {

  # check if data is a dataframe or tibble
  if (!('data.frame' %in% class(data))) {
    stop("Input data type is not of class data.frame.")
  }
  #check that dataframe is not empty
  if (nrow(data) == 0){
    stop("Cannot summarize a dataframe without observations")
  }
  #for each column do numerical or categorical summary on the column data
  cols <- colnames(data)
  all_stats <- list()
  for (i in cols) {
    vec <- get(i, data)
    if(is.numeric(vec)) {
      stats <- list(
        "count"       =   length(vec),
        "count_unique"=   length(unique(vec)),
        "unique"      =   unique(vec),
        "count_NAs"   =   sum(is.na(vec)),
        "min_"        =   min(vec, na.rm = T),
        "max_"        =   max(vec, na.rm = T),
        "mean"        =   mean(vec, na.rm = T),
        "median"      =   median(vec, na.rm = T)
      )
      names(stats) <- c("count", "count_unique", "unique", "count_NAs", "min_", "max_", "mean", "median")
      } else if (is.character(vec)) {
      stats <- list(
        "count"        = length(vec),
        "count_unique" = length(unique(vec)),
        "unique"       = unique(vec),
        "count_NAs"    = sum(is.na(vec))
      )
      names(stats) <- c("count", "count_unique", "unique", "count_NAs")
      }
    all_stats <- append(all_stats, list(stats))
    }
    names(all_stats) <- cols
    return(all_stats)
}