Build a Team Net-Rating Dashboard Table

Net rating is the one number I'd keep if I could keep only one to judge an NBA team. It folds offense and defense into a single figure - and it predicts who's actually good far better than a won-lost record, which is noisy and schedule-dependent. So we'll go get it: every team's advanced ratings for 2023-24, ranked into a clean dashboard table, then a diverging bar chart that splits the league into the good, the bad, and the merely average.

This builds on your first NBA pull - same headers, same fallback pattern, a different endpoint. If you haven't set up nba_api yet, start there; here we assume it's installed and reuse the shared NBA_HEADERS.

As with every NBA tutorial, the live pull below works on your home connection but times out from data-center and VPN IPs, so this page was built from a bundled real sample - team ratings from Basketball-Reference. The numbers are genuine; on your machine the same code fetches them live.

1. What net rating actually measures

   Net rating answers a clean question: per 100 possessions, how many more points does a team score than it allows? It's built from two halves. Offensive rating (ORtg) is points scored per 100 possessions; defensive rating (DRtg) is points allowed per 100 possessions. Net rating is simply the difference:

   net_rating = off_rating - def_rating  # ORtg - DRtg

   Why "per 100 possessions" instead of "per game"? Because teams play at different paces. A fast team and a slow team can score wildly different raw point totals while being equally efficient. Normalizing to possessions strips pace out, so a +6 net rating means the same thing for the run-and-gun Pacers as it does for a grind-it-out defense. A positive number means you outscore opponents; a negative number means you're getting outscored.

2. Pull advanced team stats

   Basic box-score endpoints won't give you ratings - you need the advanced measure type. The endpoint is LeagueDashTeamStats, and the key argument is measure_type_detailed_defense="Advanced", which switches the response from raw totals to the efficiency metrics we want. As always, we pass our browser headers and a timeout.

   from nba_api.stats.endpoints import leaguedashteamstats
   
   s = leaguedashteamstats.LeagueDashTeamStats(
       season="2023-24", measure_type_detailed_defense="Advanced",
       headers=NBA_HEADERS, timeout=30)
   df = s.get_data_frames()[0]
   
   ratings = pd.DataFrame({
       "Team": df["TEAM_NAME"], "ORtg": df["OFF_RATING"],
       "DRtg": df["DEF_RATING"], "NRtg": df["NET_RATING"],
   })

   The advanced response already includes NET_RATING, so we don't have to compute it - but it's exactly OFF_RATING - DEF_RATING, and it's a good habit to verify that yourself once with (ratings["ORtg"] - ratings["DRtg"]).round(1). The numbers should match the NRtg column to a rounding error.

3. Wrap it with the live-or-bundled fallback

   Same defensive structure as before: try the live call, and if the server blocks us, load the bundled real ratings instead. The helper retries with exponential backoff and returns a source label so the chart can credit its data honestly.

   def bundled_loader():
       return sdt_nba.bundled("nba_ratings.csv")[["Team", "ORtg", "DRtg", "NRtg"]]
   
   ratings, source = sdt_nba.live_or_bundled(
       live_call, bundled_loader, "advanced team stats")
   ratings = ratings.sort_values("NRtg", ascending=False).reset_index(drop=True)

   Sorting by NRtg descending puts the best team first. The reset_index(drop=True) renumbers the rows cleanly so we can build a 1-based rank in the next step. If you're running at home, live_call never fails and the bundled loader is simply never reached.

4. Build the ranked dashboard table

   Now we turn the DataFrame into something that reads like a real leaderboard: a 1-based rank index, every column rounded to one decimal, the whole league from best to worst.

   show = ratings.round(1).copy()
   show.index = range(1, len(show) + 1)
   print("Team net-rating table, 2023-24 (best to worst):")
   print(show.to_string())

   Full net-rating table, 2023-24

   Team net-rating table, 2023-24 (best to worst):
                         Team   ORtg   DRtg  NRtg
   1           Boston Celtics  124.2  112.5  11.7
   2    Oklahoma City Thunder  120.4  113.1   7.3
   3   Minnesota Timberwolves  116.6  110.0   6.6
   4           Denver Nuggets  119.5  114.0   5.4
   5          New York Knicks  119.1  114.2   4.9
   6     New Orleans Pelicans  118.3  113.7   4.6
   7     Los Angeles Clippers  119.8  116.5   3.2
   8             Phoenix Suns  118.8  115.7   3.1
   9       Philadelphia 76ers  117.8  114.8   3.0
   10          Indiana Pacers  121.8  118.8   3.0
   11   Golden State Warriors  118.7  116.0   2.7
   12         Milwaukee Bucks  119.1  116.5   2.6
   13     Cleveland Cavaliers  116.2  113.8   2.4
   14           Orlando Magic  114.4  112.3   2.1
   15        Dallas Mavericks  118.4  116.3   2.1
   16        Sacramento Kings  117.7  116.0   1.7
   17              Miami Heat  115.0  113.3   1.6
   18         Houston Rockets  115.3  114.2   1.1
   19      Los Angeles Lakers  116.7  116.2   0.5
   20           Chicago Bulls  115.8  117.5  -1.6
   21           Atlanta Hawks  118.2  120.2  -2.1
   22           Brooklyn Nets  114.1  117.1  -3.0
   23               Utah Jazz  116.4  121.4  -5.0
   24       San Antonio Spurs  110.8  117.2  -6.3
   25         Toronto Raptors  113.1  119.7  -6.5
   26       Memphis Grizzlies  108.1  115.3  -7.2
   27         Detroit Pistons  110.4  119.5  -9.1
   28      Washington Wizards  111.1  120.2  -9.1
   29  Portland Trail Blazers  109.3  118.6  -9.3
   30       Charlotte Hornets  110.3  120.9 -10.6

   Read it top to bottom. The Boston Celtics weren't just the best team - they were in a tier of their own at +11.7, built on a 124.2 offense and a 112.5 defense, both elite. Oklahoma City is a distant second at +7.3. Down at the bottom, Charlotte sits at -10.6, scoring 110.3 and bleeding 120.9 the other way. The whole spread, from +11.7 to -10.6, is what separates a 64-win juggernaut from a lottery team.

   Two teams worth a second look: Indiana posted a sky-high 121.8 offense but a leaky 118.8 defense, netting just +3.0 - elite scoring dragged back to mid-pack by defense. Memphis is the mirror image of bad luck, with the league's weakest offense at 108.1. Net rating makes those trade-offs visible in a way win totals never could.

5. Draw the diverging "dashboard" chart

   A diverging bar chart is the perfect picture of net rating because the metric has a natural zero. We color bars green above the line (winning teams) and red below it (losing teams), draw a vertical axis at zero, and label each bar with a signed value so the sign is unmistakable.

   import matplotlib.pyplot as plt
   
   plot_df = ratings.sort_values("NRtg")
   colors = [sdt.sport_color("soccer") if v >= 0 else sdt.SPORT_COLORS["baseball"]
             for v in plot_df["NRtg"]]
   fig, ax = plt.subplots(figsize=(8, 8.4))
   ax.barh(plot_df["Team"], plot_df["NRtg"], color=colors)
   ax.axvline(0, color="#20242B", linewidth=0.8)
   ax.bar_label(ax.containers[0], fmt="%+.1f", padding=3, fontsize=8)
   ax.set_title("NBA net rating by team, 2023-24")
   ax.set_xlabel("net rating (points per 100 possessions)")
   ax.tick_params(axis="y", labelsize=8)
   sdt.save_fig(fig, "net_rating_table", source=source)

   

   The list comprehension that builds colors is the key trick: it walks the net-rating column and picks green or red per team. The "%+.1f" format string forces a leading + or - on every label, and axvline(0, ...) draws the dividing line the whole chart pivots around. The footer credit reads "bundled sample" here because the build server is blocked; on your home machine it will instead credit the live stats.nba.com pull.

Troubleshooting