R/data_munging.R

In jmiahjones/lunch.time: Analyze Survey Data related to Eating Trends

# Functions for loading sheets and setting column types
#' Load Sheet Privately
#' @description Loads a private sheet from google sheets using your secret token.
#'
#' @param worksheet.title : A character string representing the title of the workbook.
#' @param sheet.index : An integer representing which worksheet to load.
#' @param format.changed : A boolean representing whether the format of the workbook
#' has been changed from the default format.
#'
#' @return Returns a data frame representing the workbook.
#' @export
#'
load_sheet <- function(worksheet.title, sheet.index, format.changed=F) {

  message(paste("Working Dir:", getwd()))

  if (file.exists(".secrets/googlesheets_token.rds")) {
    message("Token Found.")
    googlesheets::gs_auth(token=".secrets/googlesheets_token.rds", cache=F)
  } else {
    message("Token not found.")
    message(dir(".secrets"))
    # first time -- need to go through the oauth workflow
    googlesheets::gs_auth(cache=F)
    token <- googlesheets::gd_token()
    dir.create("./.secrets")
    saveRDS(token, file = "./.secrets/googlesheets_token.rds")
  }

  worksheet <- googlesheets::gs_title(worksheet.title)
  df <- googlesheets::gs_read(worksheet, ws = sheet.index, col_names=T)

  if(!format.changed)
    df <- set_columntypes(df)
  return(df)
}

#' Transform Columns
#' @description This function is called automatically by load_sheet() if format.changed
#' is false.
#'
#' This function transforms the column types of the google sheet into
#' appropriate data types. The location data is merged between the free-text "Other"
#' field and the categorical choices the user can make. The other choices are transformed
#' into factors, with the Length category transformed into an ordinal factor, increasing
#' with the time spent at lunch. The timestamp is parsed into a POSIX-style datetime
#' class, so that intuitive comparisons may be performed.
#'
#' NOTE: If your google sheet departs from the format of the original, this function
#' may introduce unexpected results. This function may be used as a model for any custom
#' column transformation. I am not liable for any headache, heartache, or frustration
#' induced by using this function on a custom sheet. You have been warned!
#'
#' @param tibble - the input data frame or tibble from the googlesheet
#'
#' @return tibble - transformed and renamed columns of the appropriate type
#' @export
#'
#' @importFrom magrittr %>%
#' @importFrom dplyr mutate
set_columntypes <- function(tibble) {

  # change column names
  df_names <- c(
    "Timestamp",
    "Lunch_Eaten",
    "Time_of_Lunch",
    "Length",
    "Location_Choice",
    "Location_Other_Answer"
  )
  colnames(tibble) <- df_names

  # now change types
  tibble <- tibble %>% mutate(
    Timestamp = as.POSIXct(Timestamp, format="%m/%d/%Y %H:%M:%S"),
    Lunch_Eaten = factor(Lunch_Eaten),
    Time_of_Lunch = factor(Time_of_Lunch),
    Full_Location =
      as.factor(
        ifelse(Location_Choice != "Other",
          Location_Choice,
          Location_Other_Answer)
      ),
    Location_Choice = factor(Location_Choice),
    Length = factor(Length, ordered=T,
      levels=c("0-15 mins","15-30 mins","30-45 mins",
               "45-60 mins","More than 1 hr")
      )
  )

  return(tibble)
}