Exporting event data to a dataframe¶

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

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

from kloppy import statsbomb # Load Statsbomb open dataset 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(

Test if the loading worked by printing the home and away teams.

In [2]:

# Get teams
home_team, away_team = dataset.metadata.teams
f"{home_team} vs {away_team}"

# Get teams home_team, away_team = dataset.metadata.teams f"{home_team} vs {away_team}"

Out[2]:

'Barcelona vs Deportivo Alavés'

Filter the shot events from the dataset.¶

In [3]:

# Only keep shots
shots = dataset.filter("shot")

# Only keep shots shots = dataset.filter("shot")

Convert to Polars DataFrame¶

Parameters:

• player_id: Includes the player’s unique identifier.

• Lambda function: Extracts player_name and is_goal status from each shot event.

  • player_name: Converts the player object to a string.
  • is_goal: Retrieves whether the shot resulted in a goal (True or False).

• coordinates_*: Includes all coordinate-related fields in the dataset.

• prev_pass_player: Captures the player who made the pass before the shot.

• engine="polars": Specifies the use of the Polars library for DataFrame processing. Alternatively, using engine="pandas" would convert the dataset into a Pandas DataFrame.

In [4]:

# Convert Kloppy dataset to Polars DataFrame
shots.to_df(
    "player_id",
    lambda event: {
        "player_name": str(event.player),
        "is_goal": event.result.is_success,
    },
    "coordinates_*",
    prev_pass_player=lambda event: str(event.prev("pass").player),
    engine="polars",
)

# Convert Kloppy dataset to Polars DataFrame shots.to_df( "player_id", lambda event: { "player_name": str(event.player), "is_goal": event.result.is_success, }, "coordinates_*", prev_pass_player=lambda event: str(event.prev("pass").player), engine="polars", )

Out[4]:

shape: (28, 6)

player_id	player_name	is_goal	coordinates_x	coordinates_y	prev_pass_player
str	str	bool	f64	f64	str
"5503"	"Lionel Andrés Messi Cuccittini"	false	111.45	52.85	"Ivan Rakitić"
"5211"	"Jordi Alba Ramos"	false	113.85	26.35	"Lionel Andrés Messi Cuccittini"
"5503"	"Lionel Andrés Messi Cuccittini"	false	93.65	34.65	"Rubén Duarte Sánchez"
"6613"	"Rubén Sobrino Pozuelo"	false	109.15	39.05	"Manuel Alejandro García Sánche…
"5246"	"Luis Alberto Suárez Díaz"	false	107.75	24.65	"Ousmane Dembélé"
…	…	…	…	…	…
"6935"	"Adrián Marín Gómez"	false	114.45	32.75	"Ibai Gómez Pérez"
"3501"	"Philippe Coutinho Correia"	false	113.15	31.35	"Lionel Andrés Messi Cuccittini"
"3501"	"Philippe Coutinho Correia"	true	105.25	33.35	"Arthur Henrique Ramos de Olive…
"5503"	"Lionel Andrés Messi Cuccittini"	false	106.55	46.75	"Sergi Roberto Carnicer"
"5503"	"Lionel Andrés Messi Cuccittini"	true	111.45	36.15	"Luis Alberto Suárez Díaz"

Filter using lambda functions¶

You can also use lambda functions to apply filtering. For eg. this snippet demonstrates how to filter events using a lambda function to extract actions performed by left center-backs (LCBs) and convert them into a Polars DataFrame.

In [5]:

left_centerbacks_actions = dataset.filter(
    lambda event: event.player
    and event.player.starting_position
    and str(event.player.starting_position.code) == "LCB"
).to_df(
    "player_id",
    lambda event: {"player_name": str(event.player), "event_name": event.event_name},
    "coordinates_*",
    engine="polars",
)

left_centerbacks_actions

left_centerbacks_actions = dataset.filter( lambda event: event.player and event.player.starting_position and str(event.player.starting_position.code) == "LCB" ).to_df( "player_id", lambda event: {"player_name": str(event.player), "event_name": event.event_name}, "coordinates_*", engine="polars", ) left_centerbacks_actions

Out[5]:

shape: (252, 5)

player_id	player_name	event_name	coordinates_x	coordinates_y
str	str	str	f64	f64
"6855"	"Guillermo Alfonso Maripán Loay…	"Ball Receipt*"	33.75	27.95
"6855"	"Guillermo Alfonso Maripán Loay…	"carry"	33.75	27.95
"6855"	"Guillermo Alfonso Maripán Loay…	"pass"	36.75	27.25
"5492"	"Samuel Yves Umtiti"	"Ball Receipt*"	36.55	33.25
"5492"	"Samuel Yves Umtiti"	"carry"	36.55	33.25
…	…	…	…	…
"5492"	"Samuel Yves Umtiti"	"pass"	57.95	21.55
"6855"	"Guillermo Alfonso Maripán Loay…	"pass"	12.25	42.05
"5492"	"Samuel Yves Umtiti"	"interception"	15.65	39.65
"5492"	"Samuel Yves Umtiti"	"pass"	15.65	39.65
"6855"	"Guillermo Alfonso Maripán Loay…	"pressure"	15.45	41.35