Computing 'per 90' metrics¶

This guide explains how to process a dataset using Kloppy and Polars to analyze player passing per 90 minutes (p90). It covers extracting minutes played, filtering passes, and computing successful passes p90 and total passes p90 metrics.

Setup¶

Start by loading some event data using the Kloppy module. For the sake of this demonstration, we will use Statsbomb Open Event Data.

In [1]:

from kloppy import statsbomb
import polars as pl

dataset = statsbomb.load_open_data(
    match_id=15946,
    # Optional arguments
    coordinates="statsbomb",
)

from kloppy import statsbomb import polars as pl dataset = statsbomb.load_open_data( match_id=15946, # Optional arguments coordinates="statsbomb", )

/cw/dtaijupiter/NoCsBack/dtai/pieterr/Projects/kloppy/kloppy/_providers/statsbomb.py:83: UserWarning: 

You are about to use StatsBomb public data.
By using this data, you are agreeing to the user agreement. 
The user agreement can be found here: https://github.com/statsbomb/open-data/blob/master/LICENSE.pdf

  warnings.warn(

Extract Minutes Played¶

In [2]:

mins_played = dataset.aggregate("minutes_played")

for item in mins_played:
    print(f"{item.player} - {item.duration.total_seconds() / 60:.1f} minutes played")

mins_played = dataset.aggregate("minutes_played") for item in mins_played: print(f"{item.player} - {item.duration.total_seconds() / 60:.1f} minutes played")

Philippe Coutinho Correia - 47.5 minutes played
Sergio Busquets i Burgos - 84.3 minutes played
Jordi Alba Ramos - 92.6 minutes played
Gerard Piqué Bernabéu - 92.6 minutes played
Luis Alberto Suárez Díaz - 92.6 minutes played
Ivan Rakitić - 92.6 minutes played
Ousmane Dembélé - 76.3 minutes played
Samuel Yves Umtiti - 92.6 minutes played
Lionel Andrés Messi Cuccittini - 92.6 minutes played
Nélson Cabral Semedo - 45.1 minutes played
Sergi Roberto Carnicer - 92.6 minutes played
Arturo Erasmo Vidal Pardo - 8.3 minutes played
Arthur Henrique Ramos de Oliveira Melo - 16.3 minutes played
Marc-André ter Stegen - 92.6 minutes played
Borja González Tomás - 24.6 minutes played
Jonathan Rodríguez Menéndez - 68.0 minutes played
Rubén Duarte Sánchez - 92.6 minutes played
Rubén Sobrino Pozuelo - 70.6 minutes played
Víctor Laguardia Cisneros - 92.6 minutes played
Ibai Gómez Pérez - 92.6 minutes played
Martín Aguirregabiria Padilla - 92.6 minutes played
Jorge Franco Alviz - 22.0 minutes played
Mubarak Wakaso - 92.6 minutes played
Fernando Pacheco Flores - 92.6 minutes played
Manuel Alejandro García Sánchez - 92.6 minutes played
Daniel Alejandro Torres Rojas - 67.9 minutes played
Guillermo Alfonso Maripán Loaysa - 92.6 minutes played
Adrián Marín Gómez - 24.7 minutes played

Compute Passes Per 90 Minutes¶

First, we filter the dataset to include pass events only and convert to Polars DataFrame. Then, we calculate minutes played by each player.

In [3]:

# Only keep Passes and convert to Polars DataFrame
passes_polar = dataset.filter("pass").to_df(
    "player_id",
    lambda event: {
        "player_name": str(event.player),
        "success": event.result.is_success if event.result is not None else None,
    },
    engine="polars",
)

# Calculate minutes played
mins_played_pl = pl.DataFrame(
    [
        {
            "player_id": item.player.player_id,
            "minutes_played": item.duration.total_seconds() / 60,
        }
        for item in mins_played
    ]
)

# Only keep Passes and convert to Polars DataFrame passes_polar = dataset.filter("pass").to_df( "player_id", lambda event: { "player_name": str(event.player), "success": event.result.is_success if event.result is not None else None, }, engine="polars", ) # Calculate minutes played mins_played_pl = pl.DataFrame( [ { "player_id": item.player.player_id, "minutes_played": item.duration.total_seconds() / 60, } for item in mins_played ] )

Now, to calculate the p90 metrics:

• Group by player_id and player_name to aggregate pass statistics.
• Compute successful passes and total passes.
• Join with mins_played_pl to include minutes played.
• Calculate Per 90 Metrics:
  • success_p90: Successful passes per 90 minutes.
  • total_p90: Total passes per 90 minutes.

In [4]:

# Calculate p90 metrics
passes_p90 = (
    passes_polar.group_by("player_id", "player_name")
    .agg(successful_passes=pl.sum("success"), total_passes=pl.len())
    .join(mins_played_pl, on="player_id")
    .with_columns(
        success_p90=pl.col("successful_passes") / pl.col("minutes_played") * 90,
        total_p90=pl.col("total_passes") / pl.col("minutes_played") * 90,
    )
)

passes_p90

# Calculate p90 metrics passes_p90 = ( passes_polar.group_by("player_id", "player_name") .agg(successful_passes=pl.sum("success"), total_passes=pl.len()) .join(mins_played_pl, on="player_id") .with_columns( success_p90=pl.col("successful_passes") / pl.col("minutes_played") * 90, total_p90=pl.col("total_passes") / pl.col("minutes_played") * 90, ) ) passes_p90

Out[4]:

shape: (28, 7)

player_id	player_name	successful_passes	total_passes	minutes_played	success_p90	total_p90
str	str	u32	u32	f64	f64	f64
"3501"	"Philippe Coutinho Correia"	46	52	47.517683	87.12546	98.48965
"5203"	"Sergio Busquets i Burgos"	77	83	84.319283	82.187606	88.591835
"5211"	"Jordi Alba Ramos"	117	128	92.618717	113.691923	124.380907
"5213"	"Gerard Piqué Bernabéu"	76	81	92.618717	73.851164	78.709793
"5246"	"Luis Alberto Suárez Díaz"	28	39	92.618717	27.208323	37.897308
…	…	…	…	…	…	…
"6629"	"Fernando Pacheco Flores"	12	26	92.618717	11.66071	25.264872
"6632"	"Manuel Alejandro García Sánche…	16	21	92.618717	15.547613	20.406243
"6839"	"Daniel Alejandro Torres Rojas"	12	16	67.909867	15.903433	21.204577
"6855"	"Guillermo Alfonso Maripán Loay…	11	16	92.618717	10.688984	15.547613
"6935"	"Adrián Marín Gómez"	5	7	24.70885	18.212098	25.496937