PWHL — additional Python functions

Hand-written wrappers, loaders, and helpers in sportsdataverse.pwhl not covered by the generated API-endpoint reference above.

Dataset loaders

load_pwhl_games(return_as_pandas: 'bool' = False)

Load the PWHL games-in-data-repo manifest (no seasons argument).

Mirrors fastRhockey (R) load_pwhl_games() which reads a manifest of every PWHL game that has processed data in the data repository.

Tries the sportsdataverse-data release asset first; falls back to the raw fastRhockey-data GitHub path.

Parameters

Parameter	Type	Default	Description
return_as_pandas	bool	False	return a pandas DataFrame instead of polars.

Returns

A polars (or pandas) DataFrame of all games in the data repository.

col_name	type	description
game_id	character	Unique game identifier.
season	integer	Season year (echoed from arg).
game_date	character	Game date.
game_status	character	Game status text.
home_team	character	Home team name.
home_team_id	character	Home team identifier.
away_team	character	Away team name.
away_team_id	character	Away team identifier.
home_score	character	Home team final score.
away_score	character	Away team final score.
winner	character	Whether this competitor won the game.
venue	character	Venue where the game was played.
venue_url	character	URL for the venue.
game_type	character	Game type the row belongs to.
game_json	logical	Whether processed game JSON is available.
game_json_url	character	URL to the processed game JSON.
PBP	logical	Whether play-by-play data is available.
player_box	logical	Whether player box score data is available.
skater_box	logical	Whether skater box data is available.
goalie_box	logical	Whether goalie box data is available.
team_box	logical	Whether team box score data is available.
game_info	logical	Whether game info data is available.
game_rosters	logical	Whether game rosters data is available.
scoring_summary	logical	Whether scoring summary data is available.
penalty_summary	logical	Whether penalty summary data is available.
three_stars	logical	Whether three stars data is available.
officials	logical	Whether officials data is available.
shots_by_period	logical	Whether shots-by-period data is available.
shootout	logical	Whether shootout data is available.

Example

load_pwhl_games()

load_pwhl_goalie_box(seasons, return_as_pandas: 'bool' = False)

Alias of load_pwhl_goalie_boxscores() for naming parity with fastRhockey (R).

Parameters

Parameter	Type	Default	Description
seasons
return_as_pandas	bool	False

load_pwhl_player_box(seasons, return_as_pandas: 'bool' = False)

Alias of load_pwhl_player_boxscores() for naming parity with fastRhockey (R).

Parameters

Parameter	Type	Default	Description
seasons
return_as_pandas	bool	False

load_pwhl_schedule(seasons, return_as_pandas: 'bool' = False)

Alias of load_pwhl_schedules() for naming parity with fastRhockey (R).

Parameters

Parameter	Type	Default	Description
seasons
return_as_pandas	bool	False

load_pwhl_skater_box(seasons, return_as_pandas: 'bool' = False)

Alias of load_pwhl_skater_boxscores() for naming parity with fastRhockey (R).

Parameters

Parameter	Type	Default	Description
seasons
return_as_pandas	bool	False

load_pwhl_team_box(seasons, return_as_pandas: 'bool' = False)

Alias of load_pwhl_team_boxscores() for naming parity with fastRhockey (R).

Parameters

Parameter	Type	Default	Description
seasons
return_as_pandas	bool	False

Utilities & helpers

most_recent_pwhl_season() -> 'int'

Most-recent PWHL season as an end-year integer (max season_yr).

Other

LeagueConstants(hfa: 'float', margin_sd: 'float', avg_xgf: 'float', avg_total_goals: 'float', total_scale: 'float', shrink_k: 'float', prop_kappa: 'dict', pos_priors: 'dict', prop_team_volume_slope: 'float', in_game_wp_artifact: 'str', min_season: 'int') -> None

Fitted, league-specific constants for the NHL/PWHL prediction spine.

Parameters

Parameter	Type	Default	Description
hfa	float		home-ice edge, expected-goals units.
margin_sd	float		standard deviation of the final goal margin (deliberately WIDE for hockey).
avg_xgf	float		league mean even-strength xG-for, per game.
avg_total_goals	float		league mean total goals per game.
total_scale	float		multiplier converting rating differential to total-goals deviation.
shrink_k	float		games-played prior strength for rating shrinkage.
prop_kappa	dict		empirical-Bayes shrinkage strength per player-prop stat family.
pos_priors	dict		per-position (F/D) per-stat-family prior rates.
prop_team_volume_slope	float		game-script tilt on a player-prop projection (favored team -> fewer late shots-for). SEEDED PLACEHOLDER (~0.04), not yet fitted -- a future prop-fit task should estimate it from the realized shots-vs-exp_margin slope, mirroring how fit_props.py fits prop_kappa/pos_priors.
in_game_wp_artifact	str		filename of the bundled in-game win-probability model under sportsdataverse/nhl/models/.
min_season	int		earliest season this league's prediction spine supports.

as_of_ratings_split(df: 'pl.DataFrame', cutoff_date: '_dt.date', *, date_col: 'str' = 'date') -> 'pl.DataFrame'

Filter a frame to rows strictly before cutoff_date (the leakage boundary).

Parameters

Parameter	Type	Default	Description
df	DataFrame		a polars DataFrame with a date column.
cutoff_date	date		the game date being predicted; only strictly-earlier rows are kept.
date_col	str	'date'	name of the date column (default "date").

Returns

The subset of df with df[date_col] < cutoff_date.

Example

import datetime as dt
import polars as pl
from sportsdataverse.nhl.nhl_prediction_constants import as_of_ratings_split
df = pl.DataFrame({"date": [dt.date(2023, 1, 1), dt.date(2023, 1, 2)]})
as_of_ratings_split(df, dt.date(2023, 1, 2))

brier_score(y_true: 'np.ndarray', p_pred: 'np.ndarray') -> 'float'

Mean squared error between predicted probabilities and binary outcomes.

Parameters

Parameter	Type	Default	Description
y_true	ndarray		Array of binary outcomes (0/1).
p_pred	ndarray		Array of predicted probabilities in [0, 1].

Returns

The Brier score (0.0 is a perfect forecast).

Example

import numpy as np
from sportsdataverse._common.metrics import brier_score
brier_score(np.array([1, 0]), np.array([0.9, 0.1]))

calibration_table(y_true: 'np.ndarray', p_pred: 'np.ndarray', n_bins: 'int' = 10) -> 'pl.DataFrame'

Bucket predicted probabilities into bins and compare to actual outcome rates.

Parameters

Parameter	Type	Default	Description
y_true	ndarray		Array of binary outcomes (0/1).
p_pred	ndarray		Array of predicted probabilities in [0, 1].
n_bins	int	10	Number of equal-width probability bins.

Returns

A polars.DataFrame with columns bin_mid, mean_pred, mean_actual, n (one row per non-empty bin).

Example

import numpy as np
from sportsdataverse._common.metrics import calibration_table
calibration_table(np.array([1, 0, 1, 0]), np.array([0.9, 0.1, 0.8, 0.2]))

log_loss_score(y_true: 'np.ndarray', p_pred: 'np.ndarray', eps: 'float' = 1e-15) -> 'float'

Binary cross-entropy loss between predicted probabilities and outcomes.

Parameters

Parameter	Type	Default	Description
y_true	ndarray		Array of binary outcomes (0/1).
p_pred	ndarray		Array of predicted probabilities in [0, 1].
eps	float	1e-15	Clipping bound to avoid log(0).

Returns

The mean log loss.

Example

import numpy as np
from sportsdataverse._common.metrics import log_loss_score
log_loss_score(np.array([1, 0]), np.array([0.9, 0.1]))

mae(a: 'np.ndarray', b: 'np.ndarray') -> 'float'

Mean absolute error between two arrays.

Parameters

Parameter	Type	Default	Description
a	ndarray		First array of values.
b	ndarray		Second array of values (same length as a).

Returns

The mean absolute error.

Example

import numpy as np
from sportsdataverse._common.metrics import mae
mae(np.array([1.0, 2.0]), np.array([1.5, 2.5]))

pwhl_game_corsi(game_id: 'int', return_as_pandas: 'bool' = False) -> "'Union[pl.DataFrame, pd.DataFrame]'"

Player-level on-ice Corsi and Fenwick for a single PWHL game.

Computes shot-attempt counts for every player found on ice during a shot/blocked_shot/goal event, then joins their time-on-ice so per-60 rates are available.

Corsi/Fenwick note: the HockeyTech feed has no missed-shot event, so both metrics are proxies that count only shot + blocked_shot + goal. Every output row carries corsi_includes_missed = False.

Parameters

Parameter	Type	Default	Description
game_id	int		HockeyTech game identifier (integer or string).
return_as_pandas	bool	False	If True, return a pandas.DataFrame instead of a polars.DataFrame.

Returns

One row per on-ice player with columns: - player_id (Utf8) - corsi_for, corsi_against (Int64) - corsi_for_pct (Float64) - fenwick_for, fenwick_against (Int64) - fenwick_for_pct (Float64) - toi_seconds (Int64, from shifts; null if player not in shift data) - corsi_for_per60 (Float64) - corsi_includes_missed (Boolean, always False)

col_name	type	description
player_id	character	Unique player identifier.
corsi_for	integer	Total shot attempts (on goal, missed, and blocked) directed toward the opponent by this team while at the tracked strength in the PWHL game.
corsi_against	integer	Total shot attempts (on goal, missed, and blocked) directed against this team while at the tracked strength in the PWHL game.
corsi_for_pct	double	Share of all tracked shot attempts belonging to this team, expressed as a decimal (Corsi For percentage).
fenwick_for	integer	Unblocked shot attempts (on goal and missed only) directed toward the opponent by this team at the tracked strength in the PWHL game.
fenwick_against	integer	Unblocked shot attempts (on goal and missed only) directed against this team at the tracked strength in the PWHL game.
fenwick_for_pct	double	Share of all unblocked shot attempts belonging to this team, expressed as a decimal (Fenwick For percentage).
corsi_includes_missed	logical	Boolean flag indicating whether missed shots are included in the Corsi totals for this record.
toi_seconds	double	Total time on ice in seconds for this team at the tracked strength during the PWHL game.
corsi_for_per60	double	This team's Corsi For rate projected to a full 60 minutes of ice time.

pwhl_game_shifts(game_id: 'int', return_as_pandas: 'bool' = False) -> "'Union[pl.DataFrame, pd.DataFrame]'"

Parsed shift stints for a single PWHL game.

Calls the HockeyTech modulekit/gameshifts endpoint and returns one row per player-shift stint via ~sportsdataverse.hockeytech._parsers.parse_shifts.

Parameters

Parameter	Type	Default	Description
game_id	int		HockeyTech game identifier (integer or string).
return_as_pandas	bool	False	If True, return a pandas.DataFrame instead of a polars.DataFrame.

Returns

Columns include player_id, first_name, last_name, home, period, start_time, end_time, start_s, end_s, goal_on_shift, penalty_on_shift.

col_name	type	description
game_id	integer	Unique game identifier.
player_id	integer	Unique player identifier.
first_name	character	Player first name.
last_name	character	Player last name.
jersey_number	character	Jersey number.
home	integer	Whether the player's team was home.
period	integer	Period number.
start_time	character	Shift start time (MM:SS countdown clock).
end_time	character	Shift end time (MM:SS countdown clock).
length	character	Length of the streak in games.
start_s	integer	Shift start time in seconds elapsed from the start of the period.
end_s	double	Shift end time in seconds elapsed from the start of the period.
goal_on_shift	integer	Number of goals scored while this shift was active (0 or 1 in most cases).
penalty_on_shift	integer	Number of penalties called while this shift was active.

pwhl_game_summary(game_id: 'int') -> 'dict'

PWHL game summary — dict of frames (game/goals/penalties/shots_by_period/three_stars).

Parameters

Parameter	Type	Default	Description
game_id	int

pwhl_game_total(games: 'Any', ratings: 'Any', *, league: 'str' = 'pwhl', **kwargs: 'Any') -> 'Any'

PWHL per-game expected total goals (re-export of the expected-goals helper).

Delegates to sportsdataverse.nhl.nhl_player_props.nhl_game_total with league="pwhl" defaulted.

Parameters

Parameter	Type	Default	Description
games	Any		a schedule-shaped frame.
ratings	Any		a pwhl_team_ratings-shaped frame.
league	str	'pwhl'	league key (defaults to "pwhl").

Returns

The NHL core's game_id/exp_total frame, computed with PWHL constants.

Example

from sportsdataverse.pwhl.pwhl_player_props import pwhl_game_total
totals = pwhl_game_total(games, ratings)

pwhl_in_game_win_prob(pbp: 'Any', pregame_home_prob: 'float', *, league: 'str' = 'pwhl', **kwargs: 'Any') -> 'Any'

PWHL per-play live home win probability from the bundled in-game model.

Delegates to sportsdataverse.nhl.nhl_market.nhl_in_game_win_prob with league="pwhl" defaulted. NOTE: requires a committed pwhl_in_game_wp artifact, deferred until PWHL data lands (see module docstring); calling it before then raises a clear FileNotFoundError from the artifact loader, not a silent bad result.

Parameters

Parameter	Type	Default	Description
pbp	Any		a play-by-play frame shaped like load_nhl_pbp_full.
pregame_home_prob	float		the pregame home win probability anchor.
league	str	'pwhl'	league key (defaults to "pwhl").

Returns

The NHL core's per-play home_win_prob frame.

Example

from sportsdataverse.pwhl.pwhl_market import pwhl_in_game_win_prob
wp = pwhl_in_game_win_prob(pbp, pregame_home_prob=0.5)

pwhl_leaders(season: 'Optional[int]' = None, season_id: 'Optional[int]' = None, return_as_pandas: 'bool' = False) -> 'Any'

PWHL statistical leaders for a given season.

NOTE: the leadersExtended endpoint uses season_id (integer) to filter by season, not season (name string). The resolved integer is passed as the season_id param so historical-season requests return results.

Parameters

Parameter	Type	Default	Description
season	Optional[int]	None
season_id	Optional[int]	None
return_as_pandas	bool	False

Returns

col_name	type	description
rank	integer	Rank of the streak.
player_id	character	Unique player identifier.
jersey_number	character	Jersey number.
name	character	Team mascot name.
team_id	character	Unique team identifier.
team_name	character	Team name.
team_code	character	Team abbreviation.
team_logo	character	URL to the team logo image.
team_logo_small	character	URL of the small-format team logo image for the PWHL leader-board entry.
stat_formatted	character	Human-readable string representation of the player's leading statistic value (e.g., "22G", "0.95").
type_formatted	character	Human-readable label for the statistical category driving the leader-board ranking (e.g., "Goals", "Save Percentage").
photo	character	URL to the player photo.
photo_small	character	URL of the small-format headshot image for the PWHL leader-board player.
position	character	Player position.
division	character	Division identifier.

pwhl_player_box(game_id: 'int', return_as_pandas: 'bool' = False) -> 'Any'

PWHL player box score for a single game.

NOTE: returns an empty frame pending a captured fixture + correct endpoint wiring (A1.8 follow-up); not yet functional.

Parameters

Parameter	Type	Default	Description
game_id	int
return_as_pandas	bool	False

pwhl_player_game_log(player_id: 'int', return_as_pandas: 'bool' = False) -> 'Any'

PWHL player game-by-game log.

Parameters

Parameter	Type	Default	Description
player_id	int
return_as_pandas	bool	False

pwhl_player_info(player_id: 'int', return_as_pandas: 'bool' = False) -> 'Any'

PWHL player biographical info.

NOTE: returns an empty frame pending a captured fixture + correct endpoint wiring (A1.8 follow-up); not yet functional.

Parameters

Parameter	Type	Default	Description
player_id	int
return_as_pandas	bool	False

pwhl_player_props(seasons: 'Any', *, league: 'str' = 'pwhl', **kwargs: 'Any') -> 'Any'

PWHL empirical-Bayes shots/points player-prop projections.

Delegates to sportsdataverse.nhl.nhl_player_props.nhl_player_props with league="pwhl" defaulted.

Parameters

Parameter	Type	Default	Description
seasons	Any		an int or iterable of seasons.
league	str	'pwhl'	league key (defaults to "pwhl").

Returns

The NHL core's per-(player, game, stat) projection frame.

Example

from sportsdataverse.pwhl.pwhl_player_props import pwhl_player_props
props = pwhl_player_props(2024)

pwhl_player_search(name: 'str', return_as_pandas: 'bool' = False) -> 'Any'

Search for PWHL players by name.

Parameters

Parameter	Type	Default	Description
name	str
return_as_pandas	bool	False

Returns

col_name	type	description
person_id	character	Unique person identifier.
player_id	character	Unique player identifier.
active	character	Whether athlete is currently active.
first_name	character	Player first name.
last_name	character	Player last name.
phonetic_name	character	Phonetic spelling of the player name.
shoots	character	Shooting hand.
catches	character	Catching hand (goalies).
height	character	Player height in inches.
weight	character	Player weight in pounds.
rawbirthdate	character	Player's birth date as a raw string in the format returned by the PWHL HockeyTech API, typically YYYY-MM-DD.
birthdate	character	Date of birth.
birthtown	character	Player birth town.
birthprov	character	Player birth province/state.
birthcntry	character	Player birth country.
team_id	character	Unique team identifier.
jersey_number	character	Jersey number.
role_id	character	Numeric identifier for the player's primary positional role in the PWHL HockeyTech system (e.g., skater, goalie).
season_id	character	Season identifier.
role_name	character	Human-readable label for the player's primary positional role in the PWHL HockeyTech system (e.g., 'Forward', 'Defense', 'Goalie').
all_roles	character	Pipe- or comma-delimited string listing every positional or roster role associated with the player in the PWHL HockeyTech system.
last_team_name	character	Full name of the PWHL team on which the player most recently appeared.
last_team_code	character	Short abbreviation code for the PWHL team on which the player most recently appeared.
division	character	Division identifier.
position	character	Player position.
profile_image	character	URL of the player's official profile photograph from the PWHL HockeyTech feed.
score	character	Final score string.
last_active_date	character	ISO-formatted date string of the player's most recent recorded activity or roster transaction in the PWHL HockeyTech system.

pwhl_player_stats(player_id: 'int', return_as_pandas: 'bool' = False) -> 'Any'

PWHL player season stats across all seasons.

Parameters

Parameter	Type	Default	Description
player_id	int
return_as_pandas	bool	False

Returns

col_name	type	description
season_id	character	Season identifier.
season_name	character	Full season name (e.g., "2024-25 Regular Season").
shortname	character	Player short name.
playoff	character	Whether the row is playoff statistics.
career	character	Whether this is a career-stats season.
sopt_track_faceoffs	character	Flag indicating whether faceoff tracking is enabled for this player's statistical record in the HockeyTech system.
max_start_date	character	Latest game start date for the season.
veteran_status	character	Player veteran status.
veteran	character	Whether the player is a veteran.
jersey_number	character	Jersey number.
goals	character	Goals scored.
games_played	character	Games played.
assists	character	Assists.
points	character	Total points (goals + assists).
plus_minus	character	Plus/minus rating.
penalty_minutes	character	Penalty minutes.
power_play_goals	character	Power-play goals.
power_play_assists	character	Power-play assists.
shots	character	Shots on goal.
shootout_attempts	character	Shootout attempts.
shootout_goals	character	Shootout goals.
shootout_percentage	character	Shootout scoring percentage.
shooting_percentage	character	Shooting percentage.
shootout_winning_goals	character	Shootout game-winning goals.
points_per_game	character	Points per game.
short_handed_goals	character	Short-handed goals.
short_handed_assists	character	Short-handed assists.
game_winning_goals	character	Game-winning goals.
game_tieing_goals	character	Game-tying goals.
faceoff_wins	character	Faceoff wins.
faceoff_attempts	character	Faceoff attempts.
faceoff_pct	character	Faceoff win percentage.
hits	character	Hits.
team_name	character	Team name.
team_code	character	Team abbreviation.
team_city	character	Team city.
team_nickname	character	Team nickname.
team_id	character	Unique team identifier.
active	character	Whether athlete is currently active.
first_goals	character	First goals of a game.
insurance_goals	character	Insurance goals.
overtime_goals	character	Overtime goals.
unassisted_goals	character	Unassisted goals.
empty_net_goals	character	Empty-net goals.
penalty_minutes_per_game	character	Penalty minutes per game.
division	character	Division identifier.
ice_time	character	Total ice time.
ice_time_minutes_seconds	character	Ice time in minutes and seconds.
shots_blocked_by_player	character	Shots blocked by the player.
stat_type	character	Statistic type ("regular"/"playoff").

pwhl_player_toi(game_id: 'int', return_as_pandas: 'bool' = False) -> "'Union[pl.DataFrame, pd.DataFrame]'"

Per-player time-on-ice totals for a single PWHL game.

Fetches shifts via pwhl_game_shifts then aggregates via ~sportsdataverse.hockeytech._analytics.player_toi.

Parameters

Parameter	Type	Default	Description
game_id	int		HockeyTech game identifier (integer or string).
return_as_pandas	bool	False	If True, return a pandas.DataFrame instead of a polars.DataFrame.

Returns

One row per player with player_id, first_name, last_name, toi_seconds, num_shifts, avg_shift_s, sorted by toi_seconds descending.

col_name	type	description
player_id	integer	Unique player identifier.
first_name	character	Player first name.
last_name	character	Player last name.
toi_seconds	double	Total time on ice for the PWHL player during the game or reporting period, recorded in seconds.
num_shifts	integer	Total number of shifts the player took during the game or reporting period.
avg_shift_s	double	Average duration of a single shift for the player during the game or reporting period, measured in seconds.

pwhl_playoff_bracket(season: 'Optional[int]' = None, season_id: 'Optional[int]' = None, return_as_pandas: 'bool' = False) -> 'Any'

PWHL playoff bracket for a given season.

Parameters

Parameter	Type	Default	Description
season	Optional[int]	None
season_id	Optional[int]	None
return_as_pandas	bool	False

pwhl_predict_games(games: 'Any', ratings: 'Any', *, league: 'str' = 'pwhl', **kwargs: 'Any') -> 'Any'

PWHL vectorized pregame margin/win-prob/total (+ market edge).

Delegates to sportsdataverse.nhl.nhl_market.nhl_predict_games with league="pwhl" defaulted.

Parameters

Parameter	Type	Default	Description
games	Any		a schedule-shaped frame (game_id, home_team, away_team, neutral_site).
ratings	Any		a pwhl_team_ratings-shaped frame.
league	str	'pwhl'	league key (defaults to "pwhl").

Returns

The NHL core's per-game prediction frame, computed with PWHL constants.

Example

from sportsdataverse.pwhl.pwhl_market import pwhl_predict_games
preds = pwhl_predict_games(games, ratings)

pwhl_schedule(season: 'Optional[int]' = None, season_id: 'Optional[int]' = None, return_as_pandas: 'bool' = False) -> 'Any'

PWHL schedule — one row per game (matches fastRhockey pwhl_schedule).

Parameters

Parameter	Type	Default	Description
season	Optional[int]	None
season_id	Optional[int]	None
return_as_pandas	bool	False

Returns

col_name	type	description
game_id	character	Unique game identifier.
game_date	character	Game date.
game_status	character	Game status text.
home_team	character	Home team name.
home_team_id	character	Home team identifier.
home_score	character	Home team final score.
away_team	character	Away team name.
away_team_id	character	Away team identifier.
away_score	character	Away team final score.
venue	character	Venue where the game was played.
season_id	character	Season identifier.
game_type	character	Game type the row belongs to.

pwhl_scorebar(return_as_pandas: 'bool' = False) -> 'Any'

PWHL live scorebar (today ± 3 days).

Parameters

Parameter	Type	Default	Description
return_as_pandas	bool	False

Returns

col_name	type	description
id	character	Unique player identifier.
season_id	character	Season identifier.
league_id	character	League identifier of the team.
game_number	character	Game number within the schedule.
game_letter	character	Single-letter suffix appended to the game number to distinguish doubleheader or rescheduled games.
game_type	character	Game type the row belongs to.
quick_score	character	Compact score string summarizing the current or final score of the game (e.g., "3-2").
date	character	Game date (ISO 8601 datetime string).
flo_core_event_id	character	FloSports core event identifier linking this PWHL game to its FloSports broadcast event record.
flo_live_event_id	character	FloSports live-stream event identifier for this PWHL game.
game_date	character	Game date.
game_date_iso8601	character	Game date formatted as an ISO 8601 string (e.g., "2024-02-10T00:00:00") for timezone-aware parsing.
scheduled_time	character	Raw scheduled start time for the game as returned by the HockeyTech feed, typically in HH:MM:SS format.
scheduled_formatted_time	character	Human-readable local game start time string formatted for display (e.g., "7:00 PM ET").
timezone	character	Time zone of the transaction.
ticket_url	character	URL to the official ticketing page where fans can purchase tickets for this game.
home_id	character	Home team ESPN identifier.
home_code	character	Short team code (abbreviation) for the home team (e.g., "BOS", "MIN").
home_city	character	Hometown of the athlete.
home_nickname	character	Franchise nickname for the home team (e.g., "Fleet", "Frost").
home_long_name	character	Full name including city and franchise for the home team (e.g., "Boston Fleet").
home_division	character	Home team division.
home_goals	character	Home goals in the period.
home_audio_url	character	URL of the home-team radio or audio broadcast stream for this game.
home_video_url	character	URL of the home-team video broadcast stream for this game.
home_webcast_url	character	URL of the home-team webcast for online viewing of this game.
visitor_id	character	HockeyTech team identifier for the visiting team in this game.
visitor_code	character	Short team code (abbreviation) for the visiting team (e.g., "NYR", "OTT").
visitor_city	character	City name of the visiting team (e.g., "New York", "Ottawa").
visitor_nickname	character	Franchise nickname for the visiting team (e.g., "Charge", "Sceptres").
visitor_long_name	character	Full name including city and franchise for the visiting team (e.g., "Ottawa Charge").
visiting_division	character	Visiting team division.
visitor_goals	character	Number of goals scored by the visiting team at the current point in the game.
visitor_audio_url	character	URL of the visiting-team radio or audio broadcast stream for this game.
visitor_video_url	character	URL of the visiting-team video broadcast stream for this game.
visitor_webcast_url	character	URL of the visiting-team webcast for online viewing of this game.
period	character	Period number.
period_name_short	character	Abbreviated name of the current or final game period (e.g., "3rd", "OT").
period_name_long	character	Verbose name of the current or final game period (e.g., "Third Period", "Overtime").
game_clock	character	Game clock.
game_summary_url	character	URL to the official PWHL game summary page for this contest.
home_wins	character	Wins at home.
home_regulation_losses	character	Number of regulation losses accumulated by the home team at the time of this game in the current standings.
home_ot_losses	character	Home overtime losses.
home_shootout_losses	character	Number of shootout losses accumulated by the home team at the time of this game in the current standings.
visitor_wins	character	Number of wins accumulated by the visiting team at the time of this game in the current standings.
visitor_regulation_losses	character	Number of regulation losses accumulated by the visiting team at the time of this game in the current standings.
visitor_ot_losses	character	Number of overtime losses accumulated by the visiting team at the time of this game in the current standings.
visitor_shootout_losses	character	Number of shootout losses accumulated by the visiting team at the time of this game in the current standings.
game_status	character	Game status text.
intermission	character	Flag or string indicating whether the game is currently in an intermission period.
game_status_string	character	Short status label for the game's current state (e.g., "Final", "In Progress", "Scheduled").
game_status_string_long	character	Verbose status description for the game's current state, including period or overtime context.
ord	character	Ordinal sort key used by the HockeyTech scorebar feed to order games within a day.
venue_name	character	Name of the venue.
venue_location	character	City and/or arena name indicating the physical location where the game is played.
league_name	character	League name.
league_code	character	Short code identifying the league for this scorebar record (e.g., "PWHL").
timezone_short	character	Abbreviated timezone label for the game's scheduled start time (e.g., "ET", "CT").
home_logo	character	Home team logo URL.
visitor_logo	character	URL of the logo image for the visiting team.
flo_hockey_url	character	URL to the FloHockey streaming page for this PWHL game.
combined_client_code	character	Combined league-and-client identifier string used by the HockeyTech feed to distinguish multi-tenant deployments.

pwhl_season_id(return_as_pandas: 'bool' = False) -> 'Any'

All PWHL seasons with end-year + game-type labels (HockeyTech seasons).

Parameters

Parameter	Type	Default	Description
return_as_pandas	bool	False

Returns

col_name	type	description
season_id	integer	Season identifier.
season_name	character	Full season name (e.g., "2024-25 Regular Season").
season_short	character	Short season name.
career	character	Whether this is a career-stats season.
playoff	character	Whether the row is playoff statistics.
start_date	character	Season start date.
end_date	character	Season end date.
season_yr	integer	Year derived from the season name (concluding year).
game_type_label	character	Game type: "preseason", "regular", or "playoffs".

pwhl_skater_rapm(pbp: 'pl.DataFrame', shifts: 'pl.DataFrame', *, model_dir: "'str | None'" = None, **kwargs: 'Any') -> "'pl.DataFrame | pd.DataFrame'"

② PWHL skater xG RAPM -- shim over nhl_skater_rapm with league='pwhl'.

Guards on PWHL shift-chart coverage (see the module docstring); returns a documented empty frame + cli_warn rather than fitting a degenerate ridge when shifts is too thin.

Parameters

Parameter	Type	Default	Description
pbp	DataFrame		a PWHL play-by-play frame shaped like load_nhl_pbp_full.
shifts	DataFrame		a PWHL shift-chart frame shaped like load_nhl_shifts.
model_dir	str | None	None	booster directory passed through to the borrowed NHL xG boosters.

Returns

Same schema as nhl_skater_rapm: player_id:Int64, xg_rapm_off:Float64, xg_rapm_def:Float64, xg_rapm:Float64, toi_minutes:Float64. A zero-row frame with this schema (+ a cli_warn) when PWHL shift coverage is insufficient.

Example

from sportsdataverse.pwhl.pwhl_player_impact import pwhl_skater_rapm
rapm = pwhl_skater_rapm(pbp, shifts)

pwhl_skater_war(pbp: 'pl.DataFrame', shifts: 'pl.DataFrame', *, model_dir: "'str | None'" = None, **kwargs: 'Any') -> "'pl.DataFrame | pd.DataFrame'"

③ PWHL GAR/WAR composite -- shim over nhl_skater_war with league='pwhl'.

Guards on PWHL shift-chart coverage (see the module docstring).

Parameters

Parameter	Type	Default	Description
pbp	DataFrame		a PWHL play-by-play frame shaped like load_nhl_pbp_full.
shifts	DataFrame		a PWHL shift-chart frame shaped like load_nhl_shifts.
model_dir	str | None	None	booster directory passed through to the borrowed NHL xG boosters.

Returns

Same schema as nhl_skater_war: player_id:Int64, ev_off:Float64, ev_def:Float64, pp:Float64, pk:Float64, pens:Float64, faceoffs:Float64, gar:Float64, war:Float64. A zero-row frame with this schema (+ a cli_warn) when PWHL shift coverage is insufficient.

Example

from sportsdataverse.pwhl.pwhl_player_impact import pwhl_skater_war
war = pwhl_skater_war(pbp, shifts)

pwhl_special_teams_value(pbp: 'pl.DataFrame', shifts: 'pl.DataFrame', *, model_dir: "'str | None'" = None, **kwargs: 'Any') -> "'pl.DataFrame | pd.DataFrame'"

⑥ PWHL special-teams value -- shim over nhl_special_teams_value with league='pwhl'.

Guards on PWHL shift-chart coverage (see the module docstring).

Parameters

Parameter	Type	Default	Description
pbp	DataFrame		a PWHL play-by-play frame shaped like load_nhl_pbp_full.
shifts	DataFrame		a PWHL shift-chart frame shaped like load_nhl_shifts.
model_dir	str | None	None	booster directory passed through to the borrowed NHL xG boosters.

Returns

Same schema as nhl_special_teams_value: player_id:Int64, pp_toi_minutes:Float64, pk_toi_minutes:Float64, pp_value:Float64, pk_value:Float64. A zero-row frame with this schema (+ a cli_warn) when PWHL shift coverage is insufficient.

Example

from sportsdataverse.pwhl.pwhl_player_impact import pwhl_special_teams_value
st = pwhl_special_teams_value(pbp, shifts)

pwhl_standings(season: 'Optional[int]' = None, season_id: 'Optional[int]' = None, return_as_pandas: 'bool' = False) -> 'Any'

PWHL standings — one row per team.

Parameters

Parameter	Type	Default	Description
season	Optional[int]	None
season_id	Optional[int]	None
return_as_pandas	bool	False

Returns

col_name	type	description
team_code	character	Team abbreviation.
losses	character	Losses.
regulation_wins	character	Wins in regulation.
points	character	Total points (goals + assists).
goals_for	character	Goals for.
goals_against	character	Goals against.
non_reg_wins	character	Non-regulation wins.
non_reg_losses	character	Non-regulation losses.
games_remaining	character	Games remaining in the season.
percentage	character	Points percentage earned by the PWHL team (points divided by maximum possible points), expressed as a decimal between 0 and 1.
overall_rank	character	Overall recruit ranking (top recruits only; may be NA).
games_played	character	Games played.
team_rank	integer	Team rank in the standings.
team	character	Team name.
wins	integer	Wins.

pwhl_stats(season: 'Optional[int]' = None, season_id: 'Optional[int]' = None, position: 'str' = 'skaters', return_as_pandas: 'bool' = False) -> 'Any'

PWHL aggregate stats by season and position.

Parameters

Parameter	Type	Default	Description
season	Optional[int]	None
season_id	Optional[int]	None
position	str	'skaters'
return_as_pandas	bool	False

Returns

col_name	type	description
player_id	character	Unique player identifier.
shortname	character	Player short name.
first_name	character	Player first name.
last_name	character	Player last name.
name	character	Team mascot name.
phonetic_name	character	Phonetic spelling of the player name.
active	character	Whether athlete is currently active.
height	character	Player height in inches.
weight	character	Player weight in pounds.
last_years_club	character	Player's club in the previous season.
age	character	Player age.
shoots	character	Shooting hand.
position	character	Player position.
suspension_games_remaining	character	Suspension games remaining.
suspension_indefinite	character	Whether the suspension is indefinite.
rookie	character	Whether the player is a rookie.
veteran	character	Whether the player is a veteran.
draft_eligible	character	Whether the player is draft eligible.
jersey_number	character	Jersey number.
team_name	character	Team name.
team_code	character	Team abbreviation.
team_id	character	Unique team identifier.
division	character	Division identifier.
birthdate	character	Date of birth.
birthdate_year	character	Player birth year.
hometown	character	Prospect hometown.
homeprov	character	Player home province/state.
homecntry	character	Player home country.
birthtown	character	Player birth town.
birthprov	character	Player birth province/state.
birthcntry	character	Player birth country.
hometownprov	character	Player hometown and province/state.
homeplace	character	Player home place description.
games_played	character	Games played.
game_winning_goals	character	Game-winning goals.
game_tieing_goals	character	Game-tying goals.
first_goals	character	First goals of a game.
insurance_goals	character	Insurance goals.
unassisted_goals	character	Unassisted goals.
empty_net_goals	character	Empty-net goals.
overtime_goals	character	Overtime goals.
ice_time	character	Total ice time.
ice_time_avg	character	Average ice time.
goals	character	Goals scored.
shots	character	Shots on goal.
loose_ball_recoveries	character	Loose ball recoveries.
caused_turnovers	character	Turnovers caused.
turnovers	character	Turnovers committed.
hits	character	Hits.
shots_blocked_by_player	character	Shots blocked by the player.
ice_time_minutes_seconds	character	Ice time in minutes and seconds.
shooting_percentage	character	Shooting percentage.
assists	character	Assists.
points	character	Total points (goals + assists).
points_per_game	character	Points per game.
plus_minus	character	Plus/minus rating.
penalty_minutes	character	Penalty minutes.
penalty_minutes_per_game	character	Penalty minutes per game.
ice_time_per_game_avg	character	Average ice time per game.
hits_per_game_avg	character	Average hits per game.
minor_penalties	character	Minor penalties.
major_penalties	character	Major penalties.
power_play_goals	character	Power-play goals.
power_play_assists	character	Power-play assists.
power_play_points	character	Power play points.
short_handed_goals	character	Short-handed goals.
short_handed_assists	character	Short-handed assists.
short_handed_points	character	Short-handed points.
shootout_goals	character	Shootout goals.
shootout_attempts	character	Shootout attempts.
shootout_winning_goals	character	Shootout game-winning goals.
shootout_games_played	character	Games played that went to a shootout.
faceoff_attempts	character	Faceoff attempts.
faceoff_wins	character	Faceoff wins.
faceoff_pct	character	Faceoff win percentage.
faceoff_wa	character	Faceoff wins-to-attempts metric.
shots_on	character	Shots on goal count.
shootout_percentage	character	Shootout scoring percentage.
latest_team_id	character	Most recent team identifier.
num_teams	character	Number of teams the player has played for.
logo	character	URL to the team logo.
rank	integer	Rank of the streak.
player_page_link	character	URL to the player page.
player_image	character	URL to the player's headshot image as served by the HockeyTech/PWHL data feed.
namelink	character	HTML link for the player name.
teamlink	character	HTML link for the team.
team_breakdown	integer	Per-team statistical breakdown.
is_total	double	Whether the row is a season total.

pwhl_streaks(return_as_pandas: 'bool' = False) -> 'Any'

Current PWHL player/team streaks.

Parameters

Parameter	Type	Default	Description
return_as_pandas	bool	False

pwhl_team_ratings(seasons: 'Any', *, league: 'str' = 'pwhl', **kwargs: 'Any') -> 'Any'

PWHL opponent-adjusted, shrunk even-strength xG team ratings.

Delegates to sportsdataverse.nhl.nhl_team_ratings.nhl_team_ratings with league="pwhl" defaulted. Oracle gate deferred (no xG-bearing PWHL pbp yet -- see module docstring).

Parameters

Parameter	Type	Default	Description
seasons	Any		an int or iterable of seasons.
league	str	'pwhl'	league key (defaults to "pwhl").

Returns

The NHL core's ratings frame, computed with PWHL constants.

Example

from sportsdataverse.pwhl.pwhl_team_ratings import pwhl_team_ratings
ratings = pwhl_team_ratings(2024)

pwhl_team_roster(team_id: 'int', season: 'Optional[int]' = None, season_id: 'Optional[int]' = None, return_as_pandas: 'bool' = False) -> 'Any'

PWHL team roster for a given team + season.

Parameters

Parameter	Type	Default	Description
team_id	int
season	Optional[int]	None
season_id	Optional[int]	None
return_as_pandas	bool	False

Returns

col_name	type	description
id	character	Unique player identifier.
person_id	character	Unique person identifier.
active	character	Whether athlete is currently active.
first_name	character	Player first name.
last_name	character	Player last name.
phonetic_name	character	Phonetic spelling of the player name.
display_name	character	Player display name.
shoots	character	Shooting hand.
hometown	character	Prospect hometown.
homeprov	character	Player home province/state.
homecntry	character	Player home country.
homeplace	character	Player home place description.
birthtown	character	Player birth town.
birthprov	character	Player birth province/state.
birthcntry	character	Player birth country.
birthplace	character	City, province/state, or country where the player was born, as recorded in the PWHL HockeyTech roster.
height	character	Player height in inches.
weight	character	Player weight in pounds.
height_hyphenated	character	Player's height formatted as feet-inches with a hyphen separator (e.g., '5-9'), as supplied by the PWHL HockeyTech API.
hidden	character	Flag indicating whether the player's roster entry is suppressed from public-facing displays in the PWHL HockeyTech system.
current_team	character	Name or identifier of the PWHL team to which the player is currently rostered.
player_id	character	Unique player identifier.
status	character	Status string (e.g. captain markers).
birthdate	character	Date of birth.
birthdate_year	character	Player birth year.
rawbirthdate	character	Player's birth date as a raw string in the format returned by the PWHL HockeyTech API, typically YYYY-MM-DD.
latest_team_id	character	Most recent team identifier.
veteran_status	character	Player veteran status.
veteran_description	character	Text label or descriptor indicating the player's veteran status or experience classification in the PWHL.
team_name	character	Team name.
division	character	Division identifier.
tp_jersey_number	character	Jersey number assigned to the player on the current PWHL team roster, as provided by the HockeyTech feed.
rookie	character	Whether the player is a rookie.
position_id	character	Official position identifier.
position	character	Player position.
nhlteam	character	Name or identifier of the NHL organization that holds the player's NHL rights, if applicable.
player_id_1	character	Alternate or secondary HockeyTech player identifier, distinct from the primary person_id and player_id fields.
is_rookie	character	Whether the player is a rookie.
h	character	Hits.
w	character	Wins.
draft_status	character	Text description of the player's draft history or eligibility status (e.g., undrafted, drafted year and round).
name	character	Team mascot name.
player_image	character	URL of the player's official roster photograph from the PWHL HockeyTech feed.
catches	character	Catching hand (goalies).

pwhl_teams(season: 'Optional[int]' = None, season_id: 'Optional[int]' = None, return_as_pandas: 'bool' = False) -> 'Any'

PWHL teams for a given season.

Parameters

Parameter	Type	Default	Description
season	Optional[int]	None
season_id	Optional[int]	None
return_as_pandas	bool	False

Returns

col_name	type	description
team_name	character	Team name.
team_id	character	Unique team identifier.
team_code	character	Team abbreviation.
team_nickname	character	Team nickname.
team_label	character	Short city label.
division	character	Division identifier.
team_logo	character	URL to the team logo image.

pwhl_transactions(return_as_pandas: 'bool' = False) -> 'Any'

PWHL roster transactions.

Parameters

Parameter	Type	Default	Description
return_as_pandas	bool	False

pwhl_unit_ratings(pbp: 'pl.DataFrame', shifts: 'pl.DataFrame', *, model_dir: "'str | None'" = None, **kwargs: 'Any') -> "'pl.DataFrame | pd.DataFrame'"

⑤ PWHL line/pair ratings -- shim over nhl_unit_ratings with league='pwhl'.

Guards on PWHL shift-chart coverage (see the module docstring).

Parameters

Parameter	Type	Default	Description
pbp	DataFrame		a PWHL play-by-play frame shaped like load_nhl_pbp_full.
shifts	DataFrame		a PWHL shift-chart frame shaped like load_nhl_shifts.
model_dir	str | None	None	booster directory passed through to the borrowed NHL xG boosters.

Returns

Same schema as nhl_unit_ratings: team:Utf8, unit_ids:Utf8, unit_players:Utf8, toi_minutes:Float64, on_ice_xgf:Float64, on_ice_xga:Float64, on_ice_xgf_pct:Float64, summed_rapm:Float64, unit_value:Float64. A zero-row frame with this schema (+ a cli_warn) when PWHL shift coverage is insufficient.

Example

from sportsdataverse.pwhl.pwhl_player_impact import pwhl_unit_ratings
units = pwhl_unit_ratings(pbp, shifts)

spearman_corr(a: 'np.ndarray', b: 'np.ndarray') -> 'float'

Spearman rank correlation between two arrays.

Parameters

Parameter	Type	Default	Description
a	ndarray		First array of values.
b	ndarray		Second array of values (same length as a).

Returns

The Spearman rank correlation coefficient.

Example

import numpy as np
from sportsdataverse._common.metrics import spearman_corr
spearman_corr(np.array([1, 2, 3]), np.array([3, 1, 2]))