R/calc_plan_stats.R

In alarm-redist/alarmdata: Download, Merge, and Process Redistricting Data

calc_plan_stats <- function(plans, map, calc_polsby = FALSE, ...) {

    plans <- plans %>%
        dplyr::mutate(
            total_vap = redist::tally_var(map, .data$vap),
            plan_dev =  redist::plan_parity(map),
            comp_edge = redistmetrics::comp_frac_kept(plans = redist::pl(), map),
            ndv = redist::tally_var(map, .data$ndv),
            nrv = redist::tally_var(map, .data$nrv),
            ndshare = .data$ndv/(.data$ndv + .data$nrv),
            ...
        )


    if (calc_polsby == TRUE) {
        state <- censable::match_abb(map$state[1])
        if (length(state) != 1) {
            cli_abort(c("Column {.field state} of {.arg map} could not be matched to a single state.",
                        "x" = "Please make {.field state} column correspond to the name, abbreviation, or FIPS of one state."))
        }
        single_states_polsby <- c("AK" = 0.06574469, "DE" = 0.4595251, "ND" = 0.5142261,
                                  "SD" = 0.5576591, "VT" = 0.3692381, "WY" = 0.7721791)
        if (state %in% names(single_states_polsby)) {
            plans <- plans %>% dplyr::mutate(comp_polsby = single_states_polsby[state])
        } else {
            if (state %in% c("CA", "HI", "OR")) {
                shp <- tinytiger::tt_tracts(state = censable::match_fips(state))
            } else {
                shp <- tinytiger::tt_voting_districts(state = censable::match_fips(state))
            }

            map <- map %>%
                sf::st_drop_geometry() %>%
                dplyr::left_join(y = shp, by = c("GEOID" = "GEOID20")) %>%
                sf::st_as_sf() %>%
                sf::st_transform(crs = alarm_epsg(state))

            plans <- plans %>%
                dplyr::mutate(comp_polsby = redistmetrics::comp_polsby(plans = redist::pl(), map))
        }
    }

    tally_cols <- names(map)[c(tidyselect::eval_select(tidyselect::starts_with("pop_"), map),
                               tidyselect::eval_select(tidyselect::starts_with("vap_"), map),
                               tidyselect::eval_select(tidyselect::matches("_(dem|rep)_"), map),
                               tidyselect::eval_select(tidyselect::matches("^a[dr]v_"), map))]
    for (col in tally_cols) {
        plans <- plans |>
            dplyr::mutate({{ col }} := redist::tally_var(map, map[[col]]), .before = 'ndv')

    }

    elecs <- dplyr::select(dplyr::as_tibble(map), dplyr::contains("_dem_")) %>%
        names() %>%
        stringr::str_sub(1, 6) %>%
        unique()

    elect_tb <- lapply(elecs, function(el) {
        vote_d <- dplyr::select(dplyr::as_tibble(map),
                                dplyr::starts_with(paste0(el, "_dem_")),
                                dplyr::starts_with(paste0(el, "_rep_")))
        if (ncol(vote_d) != 2) return(dplyr::tibble())
        dvote <- dplyr::pull(vote_d, 1)
        rvote <- dplyr::pull(vote_d, 2)

        plans %>%
            dplyr::mutate(
                dem = redist::group_frac(map, dvote, dvote + rvote),
                egap = redistmetrics::part_egap(plans = redist::pl(), shp = map, rvote = rvote, dvote = dvote),
                pbias = redistmetrics::part_bias(plans = redist::pl(), shp = map, rvote = rvote, dvote = dvote)
            ) %>%
            dplyr::as_tibble() %>%
            dplyr::group_by(.data$draw) %>%
            dplyr::transmute(
                draw = .data$draw,
                district = .data$district,
                e_dvs = .data$dem,
                pr_dem = .data$dem > 0.5,
                e_dem = sum(.data$dem > 0.5, na.rm = T),
                pbias = .data$pbias[1],
                egap = .data$egap[1]
            )
    }) |>
        do.call(what = dplyr::bind_rows)

    elect_tb <- elect_tb %>%
        dplyr::group_by(.data$draw, .data$district) %>%
        dplyr::summarize(dplyr::across(dplyr::everything(), mean))
    plans <- dplyr::left_join(plans, elect_tb, by = c("draw", "district"))

    split_cols <- names(map)[tidyselect::eval_select(tidyselect::any_of(c("county", "muni")), map)]
    for (col in split_cols) {
        if (col == "county") {
            plans <- plans |>
                dplyr::mutate(county_splits = redistmetrics::splits_admin(plans = redist::pl(), map, .data$county), .before = 'ndv')
        } else if (col == "muni") {
            plans <- plans |>
                dplyr::mutate(muni_splits = redistmetrics::splits_sub_admin(plans = redist::pl(), map, .data$muni), .before = 'ndv')
        } else {
            plans <- plans |>
                dplyr::mutate("{col}_splits" := redistmetrics::splits_admin(plans = redist::pl(), map, map[[col]]), .before = 'ndv')
        }
    }

    if (!("comp_polsby" %in% names(plans))) {
        plans <- plans %>%
            dplyr::mutate(comp_polsby = NA_real_)
    }

    plans <- plans %>%
        dplyr::relocate('comp_polsby', .after = 'comp_edge')

    plans
}