R/get_senate_sessions.R

In filibustr: Data Utilities for Congressional Research

#' Start and end dates of Senate sessions
#'
#' `get_senate_sessions()` returns a tibble with the beginning (convening) and
#' ending (adjournment) dates of each legislative session of the Senate.
#'
#' The data is sourced from the official Senate website, specifically
#' <https://www.senate.gov/legislative/DatesofSessionsofCongress.htm>.
#'
#' **Senate sessions explained**
#'
#' That webpage provides this explanation of Senate sessions:
#'
#' *Prior to the 74th Congress (1935-1937), the first session of a Congress
#' officially began on March 4 of odd-numbered years and ended at midnight
#' on March 3 of odd-numbered years. Since 1935, in accordance with the
#' 20th Amendment to the Constitution, Congresses have begun and ended at noon
#' on January 3 of odd-numbered years. Each two-year Congress typically
#' includes two legislative sessions, although third or special sessions were
#' common in earlier years.*
#'
#' **The `session` column**
#'
#' The `session` column is type factor, with the following levels:
#' ```{r session levels}
#' levels(get_senate_sessions()$session)
#' # Note: That's a letter S, not a number 5!
#' ```
#'
#' The Senate has had just 2 sessions in each Congress since 1941, so if you
#' are just working with more recent data, you could convert this column to
#' numeric. However, if you are working with pre-1941 data, you will likely
#' be dealing with special sessions (denoted `"S"`), not just numbered
#' sessions.
#'
#' @returns A tibble with the `begin_date` and `adjourn_date` of each
#'  session of the Senate.
#' @export
#'
#' @importFrom rlang .data
#'
#' @examples
#' get_senate_sessions()
get_senate_sessions <- function() {
  session_dates <- read_html_table(url = paste0("https://www.senate.gov",
                                                "/legislative/DatesofSessionsofCongress.htm"),
                                   css = "#SortableData_table")

  # move row 1 to column names
  names(session_dates) <- session_dates[1,] |>
    as.character() |>
    stringr::str_replace_all(" ", "_") |>
    tolower()

  session_dates <- session_dates |>
    dplyr::slice(-1)

  # remove footnotes and newline characters from dates
  session_dates |>
    dplyr::mutate(dplyr::across(
      .cols = dplyr::ends_with("_date"),
      .fns = function(.x) {
        stringr::str_remove_all(.x,
                                pattern = paste("(?<=[:digit:]{4})[:digit:]*[:space:]",
                                                "(?<=[:digit:]{4})[:digit:]*$",
                                                "\\n", sep = "|"))
      })) |>
    # split each session into its own row
    tidyr::separate_longer_delim(cols = dplyr::ends_with("_date"),
                                 delim = stringr::regex("(?<=[:digit:]{4})(?=[:alpha:])")) |>
    dplyr::mutate(session = stringr::str_split_1(dplyr::first(.data$session), ""),
                  .by = "congress") |>
    # fix column types
    dplyr::mutate(congress = as.integer(.data$congress),
                  session = as.factor(.data$session),
                  dplyr::across(dplyr::ends_with("_date"),
                                ~ as.Date(.x, format = "%b %d, %Y")))
}