NBA

This page is regenerated every week by sdvplot's docs workflow. It builds the standings with offensive, defensive and net ratings, charts net rating with logos and ranks the scoring leaders with their headshots, for the latest NBA season with games: the season to date from October to April, the final regular season once it is over. Data: hoopR's ESPN box scores, read from release files through sportsdataverse-py (no stats.nba.com calls).

NBA seasons are named by the year they end (2025-26 is 2026) and tip off in October, so until then the calendar points at the season that ended in June. For a season that is not published yet load_nba_team_boxscore warns and returns an empty frame rather than raising, so the helper below turns "no regular-season games" into a NoDataError and the page steps back one season.

import datetime as dt
import warnings

import matplotlib.pyplot as plt
import polars as pl
import sportsdataverse.nba as nba
from IPython.display import Markdown, display
from sportsdataverse.errors import NoDataError

import sdvplot

today = dt.date.today()
current = today.year + 1 if today.month >= 10 else today.year
SOURCE = "Data: hoopR (ESPN) via sportsdataverse-py"


def label(season):
    return f"{season - 1}-{season % 100:02d}"


def team_games(season):
    with warnings.catch_warnings():
        warnings.simplefilter("ignore")  # "no data for season(s)": handled just below
        box = nba.load_nba_team_boxscore(seasons=[season])
    if box.is_empty() or box.filter(pl.col("season_type") == 2).is_empty():
        raise NoDataError(f"no {label(season)} regular-season games yet")
    return box


try:
    season, box = current, team_games(current)
except NoDataError as err:
    print(f"{err}; showing {label(current - 1)} instead")
    season, box = current - 1, team_games(current - 1)

no 2026-27 regular-season games yet; showing 2025-26 instead

The status line is written when the page runs: the season to date, a finished regular season with the playoffs under way, or the offseason.

last_game = box["game_date"].max()
if season < current:
    status = f"**Offseason:** the final {label(season)} regular season; the {label(current)} season has no games yet."
    through = "final regular season"
elif box.filter(pl.col("season_type") == 3).is_empty():
    status = f"**Updated {today}:** the {label(season)} season through {last_game:%B} {last_game.day}."
    through = f"through {last_game:%b} {last_game.day}"
elif (today - last_game).days <= 10:
    status = f"**Updated {today}:** the final {label(season)} regular season; the playoffs are under way."
    through = "final regular season"
else:
    status = f"**Offseason:** the final {label(season)} regular season."
    through = "final regular season"
subtitle = through[:1].upper() + through[1:]
display(Markdown(status))

Offseason: the final 2025-26 regular season; the 2026-27 season has no games yet.

ESPN files the All-Star Game and the in-season cup final as regular-season games, though neither counts in the standings. The schedule marks them in type_abbreviation (ALLSTAR, CC), so keep the standard games (STD) only. resolve then maps the ESPN team ids to sdvplot's, for the team table's names and conferences.

standard = nba.load_nba_schedule(seasons=[season]).filter(
    (pl.col("season_type") == 2) & (pl.col("type_abbreviation") == "STD")
)
assert box.schema["game_id"] == standard.schema["game_id"]
regular = box.filter(pl.col("season_type") == 2).join(standard.select("game_id"), on="game_id", how="semi")
regular = regular.with_columns(
    team=pl.Series(sdvplot.resolve(regular["team_id"].cast(pl.String).to_list(), "nba"), dtype=pl.String)
)
possessions = (  # the box-score estimate, averaged with the opponent's
    pl.col("field_goals_attempted")
    - pl.col("offensive_rebounds")
    + pl.col("total_turnovers")
    + 0.44 * pl.col("free_throws_attempted")
)
games = regular.with_columns(poss=possessions).with_columns(poss=pl.col("poss").mean().over("game_id"))
teams = sdvplot.teams("nba").select(team="team_id", abbr="abbr", name="short_name", conference="conference")
ratings = (
    games.sort("game_date")
    .group_by("team")
    .agg(
        w=pl.col("team_winner").sum(),
        l=(~pl.col("team_winner")).sum(),
        l10=pl.col("team_winner").tail(10).sum(),
        ortg=100 * pl.col("team_score").sum() / pl.col("poss").sum(),
        drtg=100 * pl.col("opponent_team_score").sum() / pl.col("poss").sum(),
    )
    .with_columns(pct=pl.col("w") / (pl.col("w") + pl.col("l")), net=pl.col("ortg") - pl.col("drtg"))
    .join(teams, on="team")
    .sort(["pct", "net", "team"], descending=[True, True, False])  # a tiebreaker keeps re-renders stable
)
ratings.head()

team	w	l	l10	ortg	drtg	pct	net	abbr	name	conference
25	64	18	7	115.988049	105.126054	0.780488	10.861996	OKC	Thunder	Western Conference
24	62	20	8	116.606658	108.525126	0.756098	8.081532	SA	Spurs	Western Conference
8	60	22	8	114.414822	106.488601	0.731707	7.926221	DET	Pistons	Eastern Conference
2	56	26	8	117.222701	109.368854	0.682927	7.853846	BOS	Celtics	Eastern Conference
7	54	28	10	119.578358	114.537192	0.658537	5.041166	DEN	Nuggets	Western Conference

1. Standings with net rating

Each conference ordered by winning percentage, with every team's points scored and allowed per 100 possessions. gt_sdv_logos draws the logos from the abbreviations in sdvplot's team table.

from great_tables import GT

from sdvplot.great_tables import gt_save_crop, gt_sdv_logos, gt_theme_sofa

table = ratings.with_columns(
    record=pl.format("{}-{}", "w", "l"),
    last10=pl.format("{}-{}", "l10", pl.min_horizontal(10, pl.col("w") + pl.col("l")) - pl.col("l10")),
    seed=pl.col("pct").rank("ordinal", descending=True).over("conference"),
).sort("conference", "seed")
table = table.select("conference", "seed", "abbr", "name", "record", "pct", "last10", "ortg", "drtg", "net")
gt = (
    GT(table, groupname_col="conference", id="nba-standings")  # fixed id: no random one each run
    .tab_header(f"NBA standings and ratings, {label(season)}", subtitle)
    .fmt_number("pct", decimals=3)
    .cols_align("left", "name")
    .fmt_number(["ortg", "drtg"], decimals=1)
    .fmt_number("net", decimals=1, force_sign=True)
    .data_color("net", palette=["#c84630", "#f7f7f7", "#2e8b57"], domain=[-15, 15])
    .tab_spanner("Per 100 possessions", ["ortg", "drtg", "net"])
    .cols_label(
        seed="", abbr="", name="Team", record="W-L", pct="Pct", last10="Last 10", ortg="Off", drtg="Def", net="Net"
    )
    .tab_source_note(SOURCE + ". Possessions estimated from the box score; ties ordered by net rating.")
)
gt = gt_theme_sofa(gt_sdv_logos(gt, "abbr", league="nba", height=26))
gt

gt_save_crop renders the same table to a trimmed PNG, ready to post.

gt_save_crop(gt, width=900)

png

2. Net rating, best to worst

The same net ratings as bars in team colors; axis_logos swaps the abbreviations on the x axis for logos.

by_net = ratings.sort(["net", "team"], descending=[True, False])
fig, ax = plt.subplots(figsize=(10, 5.5))
ax.bar(by_net["abbr"], by_net["net"], color=sdvplot.team_colors(by_net["team"].to_list(), "nba"))
ax.axhline(0, color="#222222", lw=0.8)
ax.margins(x=0.01)
ax.set_ylabel("Net rating (points per 100 possessions)")
ax.spines[["top", "right"]].set_visible(False)
ax.set_title(f"NBA net rating, {label(season)} {through}", loc="left", fontweight="bold")
fig.text(0.99, 0.01, SOURCE, ha="right", fontsize=8, color="grey")
sdvplot.axis_logos(ax, "x", league="nba", season=season, height=0.06)
plt.show()

png

3. Scoring leaders with headshots

Points per game from the player box scores (standard games only, as above), for players who appeared in at least half of the most games any team has played. ESPN athlete ids feed add_headshots.

players = (
    nba.load_nba_player_boxscore(seasons=[season])
    .filter((pl.col("season_type") == 2) & ~pl.col("did_not_play") & pl.col("minutes").is_not_null())
    .join(standard.select("game_id"), on="game_id", how="semi")
)  # no All-Star or cup final
games_played = players.group_by("team_id").agg(pl.col("game_id").n_unique())["game_id"].max()
leaders = (
    players.sort("game_date")
    .group_by("athlete_id")
    .agg(
        name=pl.col("athlete_display_name").last(),
        team=pl.col("team_abbreviation").last(),
        gp=pl.len(),
        ppg=pl.col("points").mean(),
    )
    .filter(pl.col("gp") >= games_played / 2)
    .sort(["ppg", "athlete_id"], descending=[True, False])
    .head(12)
    .reverse()
)

fig, ax = plt.subplots(figsize=(9, 7))
y = list(range(leaders.height))
ax.barh(y, leaders["ppg"], color=sdvplot.team_colors(leaders["team"].to_list(), "nba"), height=0.7)
ax.set_yticks(y, [f"{name}  " for name in leaders["name"]])
top = leaders["ppg"].max()
for i, (ppg, gp) in enumerate(zip(leaders["ppg"], leaders["gp"], strict=True)):
    ax.text(ppg + 0.1 * top, i, f"{ppg:.1f}  ({gp} games)", va="center", fontsize=9)
ax.set_xlim(-0.11 * top, 1.42 * top)
sdvplot.add_headshots(ax, [-0.055 * top] * leaders.height, y, leaders["athlete_id"], league="nba", height=0.085)
sdvplot.add_logos(
    ax, (leaders["ppg"] + 0.05 * top).to_list(), y, leaders["team"], league="nba", season=season, height=0.06
)
ax.spines[["top", "right", "left"]].set_visible(False)
ax.tick_params(axis="y", length=0)
ax.set_xlabel("Points per game")
ax.set_title(f"NBA scoring leaders, {label(season)} {through}", loc="left", fontweight="bold")
fig.text(0.99, 0.01, f"Minimum {games_played / 2:.0f} games. {SOURCE}", ha="right", fontsize=8, color="grey")
plt.show()

png

Run it yourself

Download the notebook (outputs cleared) or open it on GitHub.

1. Standings with net rating​

2. Net rating, best to worst​

3. Scoring leaders with headshots​

Run it yourself​

1. Standings with net rating

2. Net rating, best to worst

3. Scoring leaders with headshots

Run it yourself