Getting Started with wehoop
Saiem Gilani
Source: vignettes/getting-started-wehoop.Rmd
getting-started-wehoop.RmdWelcome folks,
I’m Saiem Gilani, one of the authors of wehoop, and
I built it to give the women’s basketball community a serious,
well-tested R toolkit for play-by-play, box score, and reference data –
the same kind of analytical foundation our colleagues on the men’s side
have had for years through hoopR and cfbfastR.
This vignette walks you from a clean install through the most common
kinds of analysis you can run in your first hour with the package.
If you’ve never opened R before, the next section will get you set
up. If you already have R, RStudio, and the package installed, jump to
What’s in wehoop.
What you’ll need
wehoop runs on R 4.1.0 or newer. Most data pulls work
fine on a laptop – the heaviest single call (load_wbb_pbp()
for every season ESPN tracks) returns roughly 7 million rows and uses
about 1-2 GB of memory. If you’re on something memory-constrained,
restrict to a few seasons via seasons = 2023:2026.
Installing R and RStudio
If you’re starting from zero:
- Head to https://cran.r-project.org.
- Pick the link for your operating system:
- Windows – choose “base”, then download the most recent installer.
- macOS – pick the Latest Release. If your Mac is on an older OS version, scroll down to “Binaries for Legacy macOS Systems”.
- Linux – pick your distro and follow the install instructions.
- Then grab RStudio from Posit and follow its installer.
- The RStudio IDE Cheatsheet is worth printing – it’s one page, and it covers the keyboard shortcuts you’ll use every day.
- Windows users: also install Rtools. It’s not an R package – it’s the C/C++ toolchain R uses to build other packages from source. You’ll need it eventually even if you don’t think you do today.
Installing wehoop
You can pull a CRAN release with
install.packages("wehoop"), but most readers will want the
development version, which is on a faster release cycle than CRAN.
# pak handles dependency resolution well across OSes
if (!requireNamespace("pak", quietly = TRUE)) install.packages("pak")
pak::pkg_install(c("wehoop", "dplyr", "glue", "progressr", "tictoc"))
# Or for the development version straight from GitHub:
# pak::pak("sportsdataverse/wehoop")The four other packages above (dplyr, glue,
progressr, tictoc) aren’t strictly required by
wehoop, but every example in this vignette uses one of
them.
What’s in wehoop
It helps to know the layout of the package up front.
wehoop wraps three different upstream data
sources, and the function name tells you which one you’re
hitting:
| Prefix | Source | Best for |
|---|---|---|
wnba_* |
WNBA Stats API (stats.wnba.com) | Deep WNBA-only stats: hustle, lineups, shot charts, draft combine, league dashboards. |
espn_wnba_* |
ESPN’s WNBA endpoints | Tidy WNBA play-by-play, box scores, schedules, rosters, news. |
espn_wbb_* |
ESPN’s women’s college endpoints | The same shape of data for NCAA Division I women’s basketball. |
ncaa_wbb_* |
NCAA.com | NCAA-specific reference (NET rankings, conference standings). |
load_* |
sportsdataverse releases bucket | Pre-aggregated, multi-season parquet/RDS files. The fastest path to “every play, every season”. |
The naming is consistent enough that once you know the prefix, you
can usually guess the function. espn_wbb_pbp(game_id) and
espn_wnba_pbp(game_id) work the same way; so do their
_team_box() and _player_box() siblings.
wnba_* functions are deeper but are tied to ESPN’s older,
less-tidied API surface and tend to require parameter tuning.
The load_* functions are the ones you’ll reach for most.
They don’t hit the live API at all – they download nightly-built parquet
files from a public release bucket. That’s why they pull millions of
rows in seconds rather than the hours a per-game scrape would take.
A first-hour tour
Let’s pull every season of WNBA and WBB play-by-play that exists, plus the box scores. Each call below should land in well under a minute on a typical broadband connection.
Every WNBA play, 2002 onward
tictoc::tic()
progressr::with_progress({
wnba_pbp <- wehoop::load_wnba_pbp()
})
tictoc::toc()## 0.407 sec elapsed
## 13.91 sec elapsed
glue::glue(
"{nrow(wnba_pbp)} rows of WNBA play-by-play data from ",
"{length(unique(wnba_pbp$game_id))} games."
)## 22967 rows of WNBA play-by-play data from 55 games.
## 1782985 rows of WNBA play-by-play data from 4674 games.
dplyr::glimpse(wnba_pbp)## Rows: 22,967
## Columns: 64
## $ game_play_number <int> 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12,…
## $ id <dbl> 4018569434, 4018569437, 4018569438, 40…
## $ sequence_number <int> 4, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16…
## $ type_id <int> 615, 131, 155, 130, 155, 90, 92, 155, …
## $ type_text <chr> "Jumpball", "Pullup Jump Shot", "Defen…
## $ text <chr> "NaLyssa Smith vs. Jessica Shepard (Ja…
## $ away_score <int> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
## $ home_score <int> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3,…
## $ period_number <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,…
## $ period_display_value <chr> "1st Quarter", "1st Quarter", "1st Qua…
## $ clock_display_value <chr> "10:00", "9:40", "9:38", "9:21", "9:18…
## $ scoring_play <lgl> FALSE, FALSE, FALSE, FALSE, FALSE, FAL…
## $ score_value <int> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3,…
## $ team_id <int> 17, 17, 3, 3, 17, 17, 3, 17, 17, 17, 3…
## $ athlete_id_1 <int> 4398776, 4065870, 4399415, 3906949, 43…
## $ athlete_id_2 <int> 3906949, NA, NA, NA, NA, NA, NA, NA, N…
## $ athlete_id_3 <int> 4065870, NA, NA, NA, NA, NA, NA, NA, N…
## $ wallclock <chr> "2026-05-29T00:06:21Z", "2026-05-29T00…
## $ shooting_play <lgl> FALSE, TRUE, FALSE, TRUE, FALSE, FALSE…
## $ coordinate_x_raw <dbl> -214748340, 39, 39, 21, 21, 39, 40, 40…
## $ coordinate_y_raw <dbl> -214748365, 2, 2, 12, 12, 21, 21, 21, …
## $ points_attempted <int> 0, 2, 0, 2, 0, 0, 3, 0, 3, 0, 0, 0, 3,…
## $ short_description <chr> "Jump Ball", "Missed FG", "Rebound", "…
## $ game_id <int> 401856943, 401856943, 401856943, 40185…
## $ season <int> 2026, 2026, 2026, 2026, 2026, 2026, 20…
## $ season_type <int> 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,…
## $ home_team_id <int> 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,…
## $ home_team_name <chr> "Dallas", "Dallas", "Dallas", "Dallas"…
## $ home_team_mascot <chr> "Wings", "Wings", "Wings", "Wings", "W…
## $ home_team_abbrev <chr> "DAL", "DAL", "DAL", "DAL", "DAL", "DA…
## $ home_team_name_alt <chr> "Dallas", "Dallas", "Dallas", "Dallas"…
## $ away_team_id <int> 17, 17, 17, 17, 17, 17, 17, 17, 17, 17…
## $ away_team_name <chr> "Las Vegas", "Las Vegas", "Las Vegas",…
## $ away_team_mascot <chr> "Aces", "Aces", "Aces", "Aces", "Aces"…
## $ away_team_abbrev <chr> "LV", "LV", "LV", "LV", "LV", "LV", "L…
## $ away_team_name_alt <chr> "Las Vegas", "Las Vegas", "Las Vegas",…
## $ game_spread <dbl> 2.5, 2.5, 2.5, 2.5, 2.5, 2.5, 2.5, 2.5…
## $ home_favorite <lgl> TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TR…
## $ game_spread_available <lgl> FALSE, FALSE, FALSE, FALSE, FALSE, FAL…
## $ home_team_spread <dbl> 2.5, 2.5, 2.5, 2.5, 2.5, 2.5, 2.5, 2.5…
## $ qtr <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,…
## $ time <chr> "10:00", "9:40", "9:38", "9:21", "9:18…
## $ clock_minutes <int> 10, 9, 9, 9, 9, 9, 8, 8, 8, 8, 8, 8, 8…
## $ clock_seconds <dbl> 0, 40, 38, 21, 18, 7, 49, 47, 36, 35, …
## $ home_timeout_called <lgl> FALSE, FALSE, FALSE, FALSE, FALSE, FAL…
## $ away_timeout_called <lgl> FALSE, FALSE, FALSE, FALSE, FALSE, FAL…
## $ half <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,…
## $ game_half <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,…
## $ lead_qtr <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,…
## $ lead_half <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,…
## $ start_quarter_seconds_remaining <dbl> 600, 580, 578, 561, 558, 547, 529, 527…
## $ start_half_seconds_remaining <dbl> 1200, 1180, 1178, 1161, 1158, 1147, 11…
## $ start_game_seconds_remaining <dbl> 2400, 2380, 2378, 2361, 2358, 2347, 23…
## $ end_quarter_seconds_remaining <dbl> 600, 578, 561, 558, 547, 529, 527, 516…
## $ end_half_seconds_remaining <dbl> 1200, 1178, 1161, 1158, 1147, 1129, 11…
## $ end_game_seconds_remaining <dbl> 2400, 2378, 2361, 2358, 2347, 2329, 23…
## $ period <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,…
## $ lag_qtr <int> NA, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1…
## $ lag_half <int> NA, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1…
## $ coordinate_x <dbl> -214748406.75, -39.75, 39.75, 29.75, -…
## $ coordinate_y <dbl> -214748365, 14, -14, 4, -4, 14, -15, 1…
## $ game_date <date> 2026-05-28, 2026-05-28, 2026-05-28, 2…
## $ game_date_time <dttm> 2026-05-28 20:00:00, 2026-05-28 20:00…
## $ type_abbreviation <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA…That single tibble is the foundation for nearly any WNBA analysis
you’d want to run. The columns map cleanly onto basketball concepts:
period_number, clock_display_value,
team_id, coordinate_x,
coordinate_y, score_value,
scoring_play, and so on. Each row is one play.
A first thing you might do: count three-point attempts per team per season.
wnba_threes <- wnba_pbp %>%
dplyr::filter(
shooting_play == TRUE,
score_value %in% c(0, 3),
grepl("3", type_text, ignore.case = TRUE)
) %>%
dplyr::count(season, team_id, name = "three_attempts") %>%
dplyr::arrange(desc(season), desc(three_attempts))Or pull every shot that fell with under 5 seconds left in regulation:
clutch <- wnba_pbp %>%
dplyr::filter(
period_number == 4,
clock_minutes == 0,
clock_seconds <= 5,
scoring_play == TRUE
)The PBP table is large, so do your filtering before any grouping or summarising.
Team and player box scores
Box scores arrive at one row per (game, team) and one row per (game, player), respectively.
tictoc::tic()
progressr::with_progress({
wnba_team_box <- wehoop::load_wnba_team_box()
})
tictoc::toc()## 0.298 sec elapsed
glue::glue(
"{nrow(wnba_team_box)} rows of WNBA team boxscore data from ",
"{length(unique(wnba_team_box$game_id))} games."
)## 110 rows of WNBA team boxscore data from 55 games.
dplyr::glimpse(wnba_team_box)## Rows: 110
## Columns: 59
## $ game_id <int> 401856944, 401856944, 401856943, 401…
## $ season <int> 2026, 2026, 2026, 2026, 2026, 2026, …
## $ season_type <int> 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, …
## $ game_date <date> 2026-05-28, 2026-05-28, 2026-05-28,…
## $ game_date_time <dttm> 2026-05-28 22:00:00, 2026-05-28 22:…
## $ team_id <int> 5, 129689, 17, 3, 18, 132052, 16, 14…
## $ team_uid <chr> "s:40~l:59~t:5", "s:40~l:59~t:129689…
## $ team_slug <chr> "indiana-fever", "golden-state-valky…
## $ team_location <chr> "Indiana", "Golden State", "Las Vega…
## $ team_name <chr> "Fever", "Valkyries", "Aces", "Wings…
## $ team_abbreviation <chr> "IND", "GS", "LV", "DAL", "CON", "PO…
## $ team_display_name <chr> "Indiana Fever", "Golden State Valky…
## $ team_short_display_name <chr> "Fever", "Valkyries", "Aces", "Wings…
## $ team_color <chr> "002d62", "b38fcf", "a7a8aa", "002b5…
## $ team_alternate_color <chr> "e03a3e", "000000", "000000", "c4d60…
## $ team_logo <chr> "https://a.espncdn.com/i/teamlogos/w…
## $ team_home_away <chr> "away", "home", "away", "home", "awa…
## $ team_score <int> 88, 90, 87, 95, 61, 71, 78, 64, 81, …
## $ team_winner <lgl> FALSE, TRUE, FALSE, TRUE, FALSE, TRU…
## $ assists <int> 22, 24, 21, 23, 17, 16, 19, 14, 23, …
## $ blocks <int> 5, 11, 3, 5, 3, 3, 2, 3, 2, 2, 4, 3,…
## $ defensive_rebounds <int> 27, 29, 25, 31, 19, 20, 27, 19, 17, …
## $ fast_break_points <chr> "12", "11", "7", "4", "10", "6", "2"…
## $ field_goal_pct <dbl> 41, 46, 44, 49, 39, 46, 42, 35, 45, …
## $ field_goals_made <int> 30, 31, 36, 33, 24, 24, 28, 22, 29, …
## $ field_goals_attempted <int> 73, 68, 82, 67, 62, 52, 66, 63, 65, …
## $ flagrant_fouls <int> 0, 0, 0, 1, 0, 1, 0, 0, 1, 0, 1, 0, …
## $ fouls <int> 25, 21, 15, 16, 22, 21, 22, 21, 10, …
## $ free_throw_pct <dbl> 90, 71, 58, 86, 55, 67, 94, 79, 75, …
## $ free_throws_made <int> 19, 20, 7, 19, 12, 16, 15, 15, 12, 9…
## $ free_throws_attempted <int> 21, 28, 12, 22, 22, 24, 16, 19, 16, …
## $ largest_lead <chr> "7", "11", "13", "12", "5", "11", "2…
## $ lead_changes <chr> "9", "9", "6", "6", "13", "13", "0",…
## $ lead_percentage <chr> "15", "69", "57", "32", "21", "74", …
## $ offensive_rebounds <int> 7, 8, 8, 8, 14, 6, 13, 12, 10, 5, 10…
## $ points_in_paint <chr> "40", "42", "48", "38", "38", "32", …
## $ steals <int> 7, 5, 5, 1, 9, 10, 4, 7, 5, 10, 6, 7…
## $ team_turnovers <int> 0, 0, 0, 0, 5, 2, 0, 1, 2, 0, 1, 0, …
## $ technical_fouls <int> 0, 0, 1, 1, 1, 0, 1, 0, 0, 0, 0, 0, …
## $ three_point_field_goal_pct <dbl> 32, 38, 40, 45, 5, 32, 33, 22, 35, 3…
## $ three_point_field_goals_made <int> 9, 8, 8, 10, 1, 7, 7, 5, 11, 7, 12, …
## $ three_point_field_goals_attempted <int> 28, 21, 20, 22, 21, 22, 21, 23, 31, …
## $ total_rebounds <int> 34, 37, 33, 39, 33, 26, 40, 31, 27, …
## $ total_technical_fouls <int> 0, 0, 1, 1, 1, 0, 1, 0, 0, 0, 0, 0, …
## $ total_turnovers <int> 11, 11, 5, 10, 22, 21, 12, 12, 18, 1…
## $ turnover_points <chr> "8", "15", "7", "12", "28", "19", "8…
## $ turnovers <int> 11, 11, 5, 10, 17, 19, 12, 11, 16, 1…
## $ opponent_team_id <int> 129689, 5, 3, 17, 132052, 18, 14, 16…
## $ opponent_team_uid <chr> "s:40~l:59~t:129689", "s:40~l:59~t:5…
## $ opponent_team_slug <chr> "golden-state-valkyries", "indiana-f…
## $ opponent_team_location <chr> "Golden State", "Indiana", "Dallas",…
## $ opponent_team_name <chr> "Valkyries", "Fever", "Wings", "Aces…
## $ opponent_team_abbreviation <chr> "GS", "IND", "DAL", "LV", "POR", "CO…
## $ opponent_team_display_name <chr> "Golden State Valkyries", "Indiana F…
## $ opponent_team_short_display_name <chr> "Valkyries", "Fever", "Wings", "Aces…
## $ opponent_team_color <chr> "b38fcf", "002d62", "002b5c", "a7a8a…
## $ opponent_team_alternate_color <chr> "000000", "e03a3e", "c4d600", "00000…
## $ opponent_team_logo <chr> "https://a.espncdn.com/i/teamlogos/w…
## $ opponent_team_score <int> 90, 88, 95, 87, 71, 61, 64, 78, 96, …
tictoc::tic()
progressr::with_progress({
wnba_player_box <- wehoop::load_wnba_player_box()
})
tictoc::toc()## 0.301 sec elapsed## [1] 55
nrow(wnba_player_box)## [1] 1321A few common operations on the player box:
# Player season averages
wnba_player_box %>%
dplyr::filter(season == wehoop::most_recent_wnba_season()) %>%
dplyr::group_by(athlete_id, athlete_display_name) %>%
dplyr::summarise(
games = dplyr::n(),
ppg = mean(points, na.rm = TRUE),
rpg = mean(rebounds, na.rm = TRUE),
apg = mean(assists, na.rm = TRUE),
.groups = "drop"
) %>%
dplyr::arrange(desc(ppg))
# Career totals for one player
caitlin <- wnba_player_box %>%
dplyr::filter(athlete_display_name == "Caitlin Clark") %>%
dplyr::summarise(
games = dplyr::n(),
total_points = sum(points, na.rm = TRUE),
total_assts = sum(assists, na.rm = TRUE),
fg_pct = sum(field_goals_made, na.rm = TRUE) /
sum(field_goals_attempted, na.rm = TRUE)
)Women’s college basketball
The WBB side mirrors the WNBA side – same column shapes, same idioms, just larger volumes (Division I has ~360 programs and ~5,000 games per season).
tictoc::tic()
progressr::with_progress({
wbb_pbp <- wehoop::load_wbb_pbp()
})
tictoc::toc()## 16.444 sec elapsed## [1] 6011
nrow(wbb_pbp)## [1] 2824090
tictoc::tic()
progressr::with_progress({
wbb_team_box <- wehoop::load_wbb_team_box()
})
tictoc::toc()## 0.377 sec elapsed## [1] 6029
nrow(wbb_team_box)## [1] 12058
tictoc::tic()
progressr::with_progress({
wbb_player_box <- wehoop::load_wbb_player_box()
})
tictoc::toc()## 1.228 sec elapsed## [1] 6029
nrow(wbb_player_box)## [1] 168228If you only need a handful of seasons, both load_wbb_*()
and load_wnba_*() accept a seasons =
argument:
# Just last season's WBB box scores
recent <- wehoop::load_wbb_player_box(
seasons = (wehoop::most_recent_wbb_season() - 1):wehoop::most_recent_wbb_season()
)Bulk datasets beyond pbp and box scores
In wehoop 3.0.0 the load_*() family
expanded well past play-by-play and box scores. The same release-bucket
pattern now powers loaders for rosters, season-aggregated player and
team stats, standings, draft picks, shot events, per-game rosters, and
game officials – across all three data sources (ESPN WBB, ESPN WNBA, and
the WNBA Stats API). They all accept the same arguments as
load_wnba_pbp(): a vector of seasons =,
optional dbConnection = / tablename = for
streaming straight into a database, and progressively()
decoration of the per-season download.
A quick tour through three of them and the alternate WNBA Stats API source:
# Season-level WNBA rosters (ESPN view)
rosters_2025 <- wehoop::load_wnba_rosters(seasons = 2025)
# Per-event shots derived from PBP — every made/missed shot with court coordinates
shot_chart <- wehoop::load_wnba_shots(seasons = 2024)
# Same standings dataset, alternate source: WNBA Stats API rather than ESPN.
# Useful when you need the WNBA's official team_id keying.
stats_standings <- wehoop::load_wnba_stats_standings(seasons = 2024)
# WBB team-season stats (ESPN) for last two seasons
wbb_teams <- wehoop::load_wbb_team_stats(
seasons = (wehoop::most_recent_wbb_season() - 1):wehoop::most_recent_wbb_season()
)The ESPN-backed and WNBA Stats API-backed WNBA loaders cover overlapping ground (rosters, player stats, team stats, standings, draft, shots, game rosters, officials), so you can pick whichever joins more cleanly into the rest of your pipeline. WBB has only the ESPN-backed family today.
Live API endpoints
load_*() is the right entry point when you want history.
When you want today’s data, reach for the live
wrappers. They hit ESPN or stats.wnba.com directly, get back JSON, and
return tidy tibbles.
A few quick tasters. (None of these chunks evaluate during vignette build because they require network access; copy them into a session to run them.)
library(wehoop)
# Today's WNBA scoreboard
today_wnba <- espn_wnba_scoreboard(season = format(Sys.Date(), "%Y%m%d"))
# UConn's current roster (team_id = 2509)
uconn_roster <- espn_wbb_team_roster(team_id = 2509, season = 2025)
# A single completed game's play-by-play, team and player box, all at once
game <- espn_wnba_game_all(game_id = "401736171")
names(game) # play_by_play, Team, Player, Boxscore_team, ...
# WNBA season-leaders leaderboard
leaders <- espn_wnba_leaders(season = 2024, season_type = 2)
# Win probability per play for a single game (handy for charting momentum)
wp <- espn_wnba_event_probabilities(event_id = "401736171", limit = 200)The full ESPN surface – 80 wrappers covering rosters, schedules, news, injuries, athletes, draft, free agency, transactions, venues, coaches, and more – is documented in the ESPN basketball endpoints vignette.
For deeper WNBA stats (hustle, lineups, shot charts, draft combine),
the wnba_* family hits the WNBA Stats API directly. Those
endpoints take more parameters than the ESPN wrappers, and the parameter descriptions table is
the easiest reference for what each one accepts.
A note on rate limits and proxies
ESPN and the WNBA Stats API don’t publish official rate limits, but in practice both will return HTTP 429s or silent empty responses if you hammer them. If you’re looping over hundreds of game IDs:
- Add
Sys.sleep(1)(ESPN) orSys.sleep(3)(WNBA Stats API) between calls. - Wrap your loop in
tryCatch()so a single transient failure doesn’t halt the whole job.
If you’re behind a corporate proxy, set it once per session and every wehoop call will route through it:
options(wehoop.proxy = "http://proxy.host.example:8080")
# or, for an authenticated proxy:
options(wehoop.proxy = list(
url = "http://proxy.host.example", port = 8080,
username = "me", password = "pw", auth = "basic"
))The WNBA Stats API wrappers also accept a per-call
proxy = argument that takes precedence over the option.
Where to go from here
- ESPN basketball endpoints vignette – a guided tour of all 80 ESPN wrappers, grouped by use case.
- Parameter descriptions reference – searchable table of every WNBA Stats API query parameter.
-
wehoopreference index – every exported function, organized by data family. - The
hoopR,cfbfastR, andfastRhockeypackages share the same idioms and many of the same column conventions, if you work across multiple sports.
If you build something interesting with wehoop, please
share it – DM me on X (@saiemgilani) or open a discussion
on GitHub.
The package gets meaningfully better when users tell us what’s
missing.
