R/get_nhs_data.R

In saadaslam/fannieapi: A R interface for accessing The Exchange, which is Fannie Mae's public API

#' Get National Housing Survey data
#'
#' @return A `tbl_df` containing the following columns:
#'   * `survey_date` -- The date of the survey
#'   * `question_id` -- The ID of the question asked
#'   * `question_ans` -- The answer associated with the question in `question_id`
#'   * `percent_ans` -- The proportion of respondents who answered `question_id`
#'     as `question_ans`.
#' @export
#'
#' @examples
#' nhs_data <- get_nhs_data()
#' @seealso
#' [nhs_questions], [get_hpsi_data]
get_nhs_data <- function(){
  get <- get_url("/v1/nhs/results")

  a <- dplyr::tibble(content = get$content)

  a <- dplyr::mutate(
    a,
    names = purrr::map(content, names)
  )

  a <- tidyr::unnest(a, names, content, .id = "id1")

  a_spread <- tidyr::spread(
    a,
    names,
    content
  )

  a_spread <- tidyr::unnest(a_spread, questions, .id = "id2", .drop = FALSE)

  a_spread <- dplyr::mutate(
    a_spread,
    question_names = purrr::map(questions, names),
    date = purrr::map_chr(date, purrr::pluck),
    date = as.Date(paste0(date, "-01"), "%b-%y-%d")
  )

  b <- tidyr::unnest(
    a_spread,
    questions,
    question_names,
    .id = "id3"
  )

  b_spread <- tidyr::spread(
    b,
    question_names,
    questions
  )

  b_spread <- dplyr::mutate(
    b_spread,
    description = trimws(purrr::map_chr(description, purrr::pluck)),
    id = purrr::map_chr(id, purrr::pluck)
  )

  b_spread <- tidyr::unnest(b_spread, responses)

  b_spread <- dplyr::mutate(
    b_spread,
    response_names = purrr::map(responses, names)
  )

  b_spread <- rename(
    b_spread,
    question = description
  )

  c <- tidyr::unnest(b_spread, responses, response_names, .id = "id4")

  c_spread <- tidyr::spread(
    c,
    response_names,
    responses
  )

  c_spread <- mutate(
    c_spread,
    description = trimws(purrr::map_chr(description, purrr::pluck)),
    percent = purrr::map_dbl(percent, purrr::pluck)
  )

  result <- dplyr::select(
    c_spread,
    survey_date = date,
    question_id = id,
    question_ans = description,
    percent_ans = percent
  )

  dplyr::arrange(
    result,
    survey_date,
    question_id
  )

}