R/vms_clean.R

In CBMC-GCMP/dafishr: Download, Wrangle, and Analyse Vessel Monitoring System Data

#' Fixing dates and column names
#'
#' This function cleans raw Vessel Monitoring System, VMS, data column files,
#' eliminate NULL values in coordinates, parse dates, and returns a `data.frame`.
#'
#' It takes a raw data file downloaded using the `vms_download()` function by
#' specifying directly its path or by referencing a data.frame already stored as an R object.
#' If path is used, column with the name of the raw file is conveniently added as future reference.
#'  It also split date into three new columns `year`, `month`, `day`, and retains the original `date` column.
#' This function can be used with `apply` functions over a list
#' of files or it can be paralleled using `furrr` functions.
#'
#' @param path_to_data it can be a path to the file downloaded or the data object itself.
#' If function is used with a path it adds a `file` column to the returning data.frame object that
#' stores the name of the file as a reference.
#'
#' @return A data.frame
#' @export
#'
#' @importFrom rlang .data
#'
#' @examples
#'
#' # Using sample dataset, or a data.frame already stored as an object
#' # It is possible to use a path directly as argument
#'
#' data("sample_dataset")
#' cleaned_vms <- vms_clean(sample_dataset)
#' head(cleaned_vms)
#'
vms_clean <- function(path_to_data) {
  suppressWarnings(
    if (is.character(path_to_data) == TRUE) {
      x <- readr::read_csv(path_to_data,
        na = c("NA", "<NA>", "", " ", "?", "NULL"),
        col_types = vroom::cols(),
        locale = readr::locale(encoding = "latin1")
      )
      names(x) <- c(
        "vessel_name",
        "RNP",
        "port_base",
        "owner",
        "date",
        "latitude",
        "longitude",
        "speed",
        "direction"
      )
      empty_coordinates <- x |>
                  dplyr::filter(is.na(.data$latitude)) |>
                  dplyr::filter(is.na(.data$longitude))

      message(paste0("Cleaned: ", (nrow(empty_coordinates)), " empty rows from data! \n"))

       x  |>
        dplyr::mutate(date = as.POSIXct(date, format = "%d/%m/%Y %H:%M", tz = "UTC")) |>
        dplyr::mutate(year = lubridate::year(date)) |>
        dplyr::mutate(month = lubridate::month(date)) |>
        dplyr::mutate(day = lubridate::day(date)) |>
        dplyr::mutate(id = dplyr::row_number()) |>
        dplyr::relocate(tidyselect::any_of(c("id", "year", "month", "day", "date"))) |>
        dplyr::filter(!is.na(.data$latitude)) |>
        dplyr::filter(!is.na(.data$longitude)) |>
        dplyr::mutate(direction = as.numeric(.data$direction)) |>
        dplyr::mutate(speed = as.numeric(.data$speed)) |>
        dplyr::filter(!is.na(.data$speed)) |>
        dplyr::mutate(file_name = stringr::str_remove(path_to_data, "data/VMS-data/raw//"))



    } else if (is.data.frame(path_to_data) == TRUE) {
      x <- path_to_data

      names(x) <- c(
        "vessel_name",
        "RNP",
        "port_base",
        "owner",
        "date",
        "latitude",
        "longitude",
        "speed",
        "direction"
      )
      empty_coordinates <- x |>
                  dplyr::mutate(
                              latitude = as.numeric(.data$latitude),
                              longitude = as.numeric(.data$longitude)
                  ) |>
                  dplyr::filter(is.na(.data$latitude)) |>
                  dplyr::filter(is.na(.data$longitude))

      message(paste0("Cleaned: ", (nrow(empty_coordinates)), " empty rows from data! \n"))

      x |>
        dplyr::mutate(date = as.POSIXct(date, format = "%d/%m/%Y %H:%M", tz = "UTC")) |>
        dplyr::mutate(year = lubridate::year(date)) |>
        dplyr::mutate(month = lubridate::month(date)) |>
        dplyr::mutate(day = lubridate::day(date)) |>
        dplyr::mutate(id = dplyr::row_number()) |>
        dplyr::relocate(tidyselect::any_of(c("id", "year", "month", "day", "date"))) |>
        dplyr::mutate(
          latitude = as.numeric(.data$latitude),
          longitude = as.numeric(.data$longitude)
        ) |>
        dplyr::filter(!is.na(.data$latitude)) |>
        dplyr::filter(!is.na(.data$longitude)) |>
        dplyr::mutate(direction = as.numeric(.data$direction)) |>
        dplyr::mutate(speed = as.numeric(.data$speed))



    } else {
      stop("Data must be a path to folder or data.frame object \n")
    }
  )
}