R/Functions.R

In mahonmb/FARSfun: Maps Data from Fatality Analysis Reporting System

#' Data Input as Data Frame
#'
#' Reads a file in .csv format and creates a tibble data frame.
#'
#' @param filename A character string ending in .csv giving the file name
#'    of the data to be read into R as a data frame
#'
#' @return This function returns a dataset as a tibble data frame. As a side
#'    effect, this function prints the data frame, unless coerced as an object.
#'
#' @details If \code{filename} does not match an existing file, function will return
#'    an error: \code{"file filename does not exist"}.
#'
#' @importFrom readr read_csv
#' @importFrom dplyr tbl_df
#'
#' @examples \dontrun{fars_read(accident_2013.csv)}
#' @examples \dontrun{fars_read(accident_2014.csv)}
fars_read <- function(filename) {

  if(!file.exists(filename))
    stop("file '", filename, "' does not exist")
  data <- suppressMessages({
    readr::read_csv(filename, progress = FALSE)
  })
  dplyr::tbl_df(data)
}


#' Create File Name
#'
#' Creates a file named "accident_YEAR.csv.bz2". YEAR in the file name can be
#' specified with the \code{year} argument.
#'
#' @param year An integer or character string giving the year for the filename
#'
#' @return This function returns a file name. As a side effect, this function
#'    prints the file name.
#'
#' @examples \dontrun{make_filename(2013)}
#' @examples \dontrun{make_filename(2014)}
#'
#' @export
make_filename <- function(year) {
  year <- as.integer(year)
  sprintf("accident_%d.csv.bz2", year)
}


#' Select \code{MONTH} from Single Year Dataset
#'
#' Creates a tribble data frame containing \code{MONTH} and \code{year} of a
#' specified year's (specified with the \code{year} argument) dataset.
#'
#' @param years An integer or character string giving the year for the filename
#'
#' @return This function returns tribble data frame with column names of
#'    \code{MONTH} and \code{year}. \code{MONTH} are integers ranging from
#'    1-12 indicating January to December. As a side effect, this function
#'    returns the data frame, unless coerced as an object.
#'
#' @details If \code{year} is not contained within the datasets (2013, 2014, or 2015),
#'    the function will return a warning: \code{"invalid year:"}.
#'
#' @importFrom dplyr mutate select
#'
#' @examples \dontrun{fars_read_years(2013)}
#' @examples \dontrun{fars_read_years(2014)}
fars_read_years <- function(years) {
  lapply(years, function(year) {
    file <- make_filename(year)
    tryCatch({
      dat <- fars_read(file)
      dplyr::mutate(dat, year = year) %>%
        dplyr::select(MONTH, year)
    }, error = function(e) {
      warning("invalid year: ", year)
      return(NULL)
    })
  })
}


#' Count of Accidents for each \code{MONTH} from Single Year Dataset
#'
#' Creates a tribble data frame containing \code{MONTH} and count of observations
#' per month in a year's (specified with the \code{years} argument) dataset.
#'
#' @param years An integer or character string giving the year for the filename
#'
#' @return This function returns tribble data frame with column names of \code{MONTH}
#'    and specified \code{years}. \code{MONTH} are integers ranging from 1-12
#'    indicating January to December. The specified \code{years} column are integers
#'    representing the number of observations in each month. As a side effect, this
#'    function returns the data frame, unless coerced as an object.
#'
#' @details If \code{year} is not contained within the datasets (2013, 2014, or 2015),
#'    the function will return a warning: \code{"invalid year:"}.
#'
#' @importFrom dplyr bind_rows group_by summarize
#' @importFrom tidyr spread
#'
#' @examples \dontrun{fars_summarize_years(2013)}
#' @examples \dontrun{fars_summarize_years(2014)}
#'
#' @export
fars_summarize_years <- function(years) {
  dat_list <- fars_read_years(years)
  dplyr::bind_rows(dat_list) %>%
    dplyr::group_by(year, MONTH) %>%
    dplyr::summarize(n = n()) %>%
    tidyr::spread(year, n)
}


#' Map of Accidents within a State and Year
#'
#' This function creates a map of accident locations within the specified state
#' (\code{state.num}) and \code{year}.
#'
#' This function requires both the \code{maps} and
#' \code{graphics} packages.
#'
#' @param state.num An integer corresponding to a specific state. Indicates which
#'    state to map
#' @inheritParams make_filename
#'
#' @return This function returns a graphic plot of the accident locations within
#'    a specified state (\code{state.num}) and \code{year}. Points indicate accident
#'    locations and solid lines indicate state boundary.
#'
#' @details If \code{state.num} is not within STATE column (1-50 corresponding to 50 states
#'    in alphabetical order), will return error: \code{"invalid STATE number:"}.
#' @details If there are no accidents within the specific \code{state.num} and \code{year},
#'    function will return message: \code{"no accidents to plot"}.
#' @details If \code{year} is not contained within the datasets (2013, 2014, or 2015),
#'    the function will return a warning: \code{"invalid year:"}.
#'
#' @importFrom dplyr tbl_df
#' @importFrom maps map
#' @importFrom graphics points
#'
#' @examples \dontrun{fars_map_state(1,2013)}
#' @examples \dontrun{fars_map_state(20,2014)}
#'
#' @export
fars_map_state <- function(state.num, year) {
  filename <- make_filename(year)
  data <- fars_read(filename)
  state.num <- as.integer(state.num)

  if(!(state.num %in% unique(data$STATE)))
    stop("invalid STATE number: ", state.num)
  data.sub <- dplyr::filter(data, STATE == state.num)
  if(nrow(data.sub) == 0L) {
    message("no accidents to plot")
    return(invisible(NULL))
  }
  is.na(data.sub$LONGITUD) <- data.sub$LONGITUD > 900
  is.na(data.sub$LATITUDE) <- data.sub$LATITUDE > 90
  with(data.sub, {
    maps::map("state", ylim = range(LATITUDE, na.rm = TRUE),
              xlim = range(LONGITUD, na.rm = TRUE))
    graphics::points(LONGITUD, LATITUDE, pch = 46)
  })
}