[Football Analysis] 12.xG 집계 및 분석

작성일자 : 2023-12-28

Ver 0.1.1

 

 

슈팅 데이터 불러오기¶

In [1]:

import os
import numpy as np
import pandas as pd
import plotly.graph_objects as go

new_dir = '/Users/limjongjun/Desktop/JayJay/Growth/Python/soccer-analytics'
os.chdir(new_dir)

from src.plot_utils import get_pitch_layout

In [2]:

shots = pd.read_pickle('data/shots.pkl')
shots

Out[2]:

	competition_name	match_id	event_id	period	time	team_id	team_name	player_id	player_name	event_type	sub_event_type	tags	x	y	distance	angle	freekick	header	goal	xg
0	England	2499719	177959212	1H	94.596	1609	Arsenal	25413	A. Lacazette	Shot	Shot	[Goal, Right foot, Opportunity, Position: Goal...	12.48	6.12	13.899813	26.933236	0	0	1	0.143848
1	England	2499719	177959247	1H	179.855	1631	Leicester City	26150	R. Mahrez	Shot	Shot	[Left foot, Opportunity, Position: Out center ...	15.60	-1.36	15.659170	26.224941	0	0	0	0.120612
2	England	2499719	177959280	1H	254.745	1631	Leicester City	14763	S. Okazaki	Shot	Shot	[Goal, Head/body, Opportunity, Position: Goal ...	4.16	-1.36	4.376665	79.289489	0	1	1	0.412967
3	England	2499719	177959289	1H	425.824	1609	Arsenal	7868	A. Oxlade-Chamberlain	Shot	Shot	[Left foot, Opportunity, Position: Out high le...	19.76	11.56	22.893038	15.813597	0	0	0	0.042434
4	England	2499719	177959429	1H	815.462	1609	Arsenal	7868	A. Oxlade-Chamberlain	Shot	Shot	[Right foot, Opportunity, Position: Goal low l...	26.00	13.60	29.342120	12.655803	0	0	0	0.019414
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
45940	European_Championship	1694440	90588583	2H	2776.504	4418	France	25575	A. Gignac	Shot	Shot	[Right foot, Opportunity, Position: Post cente...	7.28	5.44	9.088014	37.600623	0	0	0	0.273871
45941	European_Championship	1694440	90589205	E1	807.318	9905	Portugal	70410	Éder	Shot	Shot	[Head/body, Opportunity, Position: Goal center...	10.40	8.16	13.219138	25.258830	0	1	0	0.056536
45942	European_Championship	1694440	90589242	E2	144.487	9905	Portugal	28907	Raphaël Guerreiro	Free kick	Free kick shot	[Left foot, Direct, Position: Post high left, ...	24.96	12.92	28.105658	13.240079	1	0	0	0.098744
45943	European_Championship	1694440	90589254	E2	204.428	9905	Portugal	70410	Éder	Shot	Shot	[Goal, Right foot, Opportunity, Position: Goal...	23.92	2.72	24.074152	17.184831	0	0	1	0.040764
45944	European_Championship	1694440	90589034	E2	981.239	4418	France	134513	A. Martial	Shot	Shot	[Right foot, Blocked, Not accurate]	13.52	6.80	15.133750	24.652934	0	0	0	0.120724

45945 rows × 20 columns

단일 경기 xG 시각화¶

(1) 슈팅 레이블 생성¶

In [3]:

match_id = 2057987

match_df = pd.read_csv('data/refined_events/World_Cup/matches.csv', index_col=0, header=0, encoding='utf-8-sig')
team1_name = match_df.at[match_id, 'team1_name']
team2_name = match_df.at[match_id, 'team2_name']

match_shots = shots[shots['match_id'] == match_id]

# mouse over 시 보일 text 
match_shots['display_name'] = match_shots.apply(
    lambda x: f"{x['player_name']}, " +
    f"{x['period']} {int(x['time'] // 60):02d}:{int(x['time'] % 60):02d}, " +
    f"xG: {round(x['xg'], 3)}", axis=1
)

team2_x = match_shots.loc[match_shots['team_name'] == team2_name, 'x']
team1_y = match_shots.loc[match_shots['team_name'] == team1_name, 'y']
team2_y = match_shots.loc[match_shots['team_name'] == team2_name, 'y']
match_shots.loc[match_shots['team_name'] == team2_name, 'x'] = 104 - team2_x
match_shots.loc[match_shots['team_name'] == team1_name, 'y'] = 34 - team1_y
match_shots.loc[match_shots['team_name'] == team2_name, 'y'] = 34 + team2_y

match_shots.head()

/var/folders/_b/znjp14gd02d8lg63thqc7bm40000gn/T/ipykernel_2682/2786941602.py:10: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  match_shots['display_name'] = match_shots.apply(

Out[3]:

	competition_name	match_id	event_id	period	time	team_id	team_name	player_id	player_name	event_type	...	tags	x	y	distance	angle	freekick	header	goal	xg	display_name
34329	World_Cup	2057987	259956999	1H	698.173	15473	Mexico	7941	J. Hernández	Shot	...	[Head/body, Opportunity, Position: Out center ...	91.52	27.88	13.899813	26.933236	0	1	0	0.054390	J. Hernández, 1H 11:38, xG: 0.054
34330	World_Cup	2057987	259957180	1H	1275.442	14855	Korea Republic	14911	Son Heung-Min	Shot	...	[Counter attack, Left foot, Blocked, Opportuni...	13.52	45.56	17.788311	18.086581	0	0	0	0.069484	Son Heung-Min, 1H 21:15, xG: 0.069
34331	World_Cup	2057987	259957182	1H	1277.147	14855	Korea Republic	14911	Son Heung-Min	Shot	...	[Right foot, Blocked, Opportunity, Not accurate]	12.48	45.56	17.011290	18.314813	0	0	0	0.075236	Son Heung-Min, 1H 21:17, xG: 0.075
34332	World_Cup	2057987	259965138	1H	1281.101	14855	Korea Republic	14911	Son Heung-Min	Shot	...	[Left foot, Blocked, Not accurate]	21.84	41.48	23.085407	17.103758	0	0	0	0.042485	Son Heung-Min, 1H 21:21, xG: 0.042
34333	World_Cup	2057987	259957220	1H	1339.263	14855	Korea Republic	61962	Sung-Yeung Ki	Shot	...	[Head/body, Opportunity, Position: Goal high c...	11.44	37.40	11.934555	32.982052	0	1	0	0.080193	Sung-Yeung Ki, 1H 22:19, xG: 0.08

5 rows × 21 columns

(2) 슈팅 위치 및 xG 시각화¶

In [4]:

match_shots_failed = match_shots[match_shots['tags'].apply(lambda x: 'Goal' not in x)]

team1_shots = match_shots[match_shots['team_name'] == team1_name]
team2_shots = match_shots[match_shots['team_name'] == team2_name]
team1_goals = team1_shots[team1_shots['tags'].apply(lambda x: 'Goal' in x)]
team2_goals = team2_shots[team2_shots['tags'].apply(lambda x: 'Goal' in x)]

team1_goal_trace = go.Scatter(
    x=team1_goals['x'],
    y=team1_goals['y'],
    name=team1_name,
    text=team1_goals['display_name'],
    mode='markers',
    marker=dict(
        color='red', size=np.sqrt(team1_goals['xg']) * 50, 
        symbol=team1_goals['freekick'].apply(lambda x: 'square' if x == 1 else 'circle')
    )
)

team2_goal_trace = go.Scatter(
    x=team2_goals['x'],
    y=team2_goals['y'],
    name=team2_name,
    text=team2_goals['display_name'],
    mode='markers',
    marker=dict(
        color='blue', size=np.sqrt(team2_goals['xg']) * 50, 
        symbol=team2_goals['freekick'].apply(lambda x: 'square' if x == 1 else 'circle')
    )
)

shot_trace = go.Scatter(
    x=match_shots_failed['x'],
    y=match_shots_failed['y'],
    name='Failed shot',
    text=match_shots_failed['display_name'],
    mode='markers',
    marker=dict(
        color='darkgrey', size=np.sqrt(match_shots_failed['xg']) * 50, 
        symbol=match_shots_failed['freekick'].apply(lambda x: 'square' if x == 1 else 'circle')
    )
)

team1_xg = team1_shots['xg'].sum().round(2)
team2_xg = team2_shots['xg'].sum().round(2)
title = f"{team1_name} - {team2_name} (xG: {team1_xg} - {team2_xg})"
fig = go.Figure(data=[shot_trace, team1_goal_trace, team2_goal_trace], layout=get_pitch_layout(title))
fig.show()

선수별 시즌 전체 xG 및 부가 지표 집계¶

(1) 선수별 슈팅/유효슈팅/득점 횟수 집계¶

In [5]:

shots = shots[shots['competition_name'] == 'England']
shots_on_target = shots[shots['tags'].apply(lambda x: 'Accurate' in x)]
goals = shots[shots['tags'].apply(lambda x: 'Goal' in x)]

player_shot_counts = shots.groupby(['team_id', 'team_name', 'player_id', 'player_name'])['event_id'].count()
player_sot_counts = shots_on_target.groupby(['team_id', 'team_name', 'player_id', 'player_name'])['event_id'].count()
player_goal_counts = goals.groupby(['team_id', 'team_name', 'player_id', 'player_name'])['event_id'].count()

player_stats = pd.concat([player_shot_counts, player_sot_counts, player_goal_counts], axis=1).fillna(0).astype(int)
player_stats.columns = ['Shots', 'SoT', 'Goals']
player_stats

Out[5]:

				Shots	SoT	Goals
team_id	team_name	player_id	player_name
1609	Arsenal	3319	M. Özil	39	14	4
		3361	A. Sánchez	69	28	7
		3560	Nacho Monreal	21	8	5
		7855	L. Koscielny	8	2	2
		7856	P. Mertesacker	2	2	1
...	...	...	...	...	...	...
10531	Swansea City	77557	W. Routledge	5	2	0
		246866	A. Mawson	12	4	2
		258162	Renato Sanches	12	0	0
		288865	O. McBurnie	3	1	0
		343951	T. Abraham	39	16	5

437 rows × 3 columns

In [6]:

player_stats.sort_values('Goals', ascending=False)[:20].reset_index()

Out[6]:

	team_id	team_name	player_id	player_name	Shots	SoT	Goals
0	1612	Liverpool	120353	Mohamed Salah	142	68	32
1	1624	Tottenham Hotspur	8717	H. Kane	175	74	29
2	1625	Manchester City	8325	S. Agüero	91	42	21
3	1631	Leicester City	12829	J. Vardy	66	34	20
4	1625	Manchester City	11066	R. Sterling	80	35	18
5	1611	Manchester United	7905	R. Lukaku	80	41	16
6	1612	Liverpool	15808	Roberto Firmino	80	37	15
7	1609	Arsenal	25413	A. Lacazette	65	36	14
8	1625	Manchester City	340386	Gabriel Jesus	55	31	13
9	1651	Brighton & Hove Albion	8416	G. Murray	50	24	12
10	1610	Chelsea	25707	E. Hazard	65	34	12
11	1624	Tottenham Hotspur	14911	Son Heung-Min	70	31	12
12	1631	Leicester City	26150	R. Mahrez	71	36	12
13	1633	West Ham United	14703	M. Arnautović	67	31	11
14	1624	Tottenham Hotspur	54	C. Eriksen	97	40	11
15	1610	Chelsea	3324	Álvaro Morata	74	33	11
16	1628	Crystal Palace	127537	L. Milivojević	42	16	10
17	1623	Everton	7944	W. Rooney	44	14	10
18	1625	Manchester City	245364	L. Sané	54	21	10
19	1612	Liverpool	25747	S. Mané	67	26	10

(2) 선수별 xG 집계 및 정렬¶

In [7]:

player_stats['xG'] = shots.groupby(['team_id', 'team_name', 'player_id', 'player_name'])['xg'].sum()
player_stats.sort_values('xG', ascending=False)[:20].reset_index()

Out[7]:

	team_id	team_name	player_id	player_name	Shots	SoT	Goals	xG
0	1624	Tottenham Hotspur	8717	H. Kane	175	74	29	25.498832
1	1612	Liverpool	120353	Mohamed Salah	142	68	32	18.787223
2	1625	Manchester City	11066	R. Sterling	80	35	18	15.184169
3	1625	Manchester City	8325	S. Agüero	91	42	21	15.125795
4	1611	Manchester United	7905	R. Lukaku	80	41	16	14.873459
5	1631	Leicester City	12829	J. Vardy	66	34	20	13.074317
6	1625	Manchester City	340386	Gabriel Jesus	55	31	13	11.803429
7	1609	Arsenal	25413	A. Lacazette	65	36	14	11.477789
8	1651	Brighton & Hove Albion	8416	G. Murray	50	24	12	10.960509
9	1633	West Ham United	14703	M. Arnautović	67	31	11	10.647402
10	1610	Chelsea	3324	Álvaro Morata	74	33	11	10.504605
11	1644	Watford	377071	Richarlison	92	23	5	10.461122
12	1628	Crystal Palace	38031	C. Benteke	55	20	3	10.098130
13	1612	Liverpool	25747	S. Mané	67	26	10	9.218263
14	1613	Newcastle United	12536	D. Gayle	53	20	6	9.031792
15	1624	Tottenham Hotspur	13484	D. Alli	70	18	9	8.882627
16	1624	Tottenham Hotspur	54	C. Eriksen	97	40	11	8.570227
17	1612	Liverpool	15808	Roberto Firmino	80	37	15	8.434821
18	1627	West Bromwich Albion	8958	J. Rodriguez	67	25	7	8.334103
19	1628	Crystal Palace	8422	W. Zaha	67	28	9	8.232400

(3) 선수별 득점-xG 차이(dG) 집계 및 정렬¶

• 높을 수록 골 결정력이 좋은 선수

In [8]:

player_stats['dG'] = player_stats['Goals'] - player_stats['xG']
player_stats.sort_values('dG', ascending=False)[:20].reset_index()

Out[8]:

	team_id	team_name	player_id	player_name	Shots	SoT	Goals	xG	dG
0	1612	Liverpool	120353	Mohamed Salah	142	68	32	18.787223	13.212777
1	1631	Leicester City	12829	J. Vardy	66	34	20	13.074317	6.925683
2	1612	Liverpool	15808	Roberto Firmino	80	37	15	8.434821	6.565179
3	1631	Leicester City	26150	R. Mahrez	71	36	12	5.961230	6.038770
4	1625	Manchester City	8325	S. Agüero	91	42	21	15.125795	5.874205
5	1624	Tottenham Hotspur	14911	Son Heung-Min	70	31	12	6.654574	5.345426
6	1625	Manchester City	245364	L. Sané	54	21	10	5.104557	4.895443
7	1610	Chelsea	25707	E. Hazard	65	34	12	7.367824	4.632176
8	1624	Tottenham Hotspur	8717	H. Kane	175	74	29	25.498832	3.501168
9	1612	Liverpool	3802	Philippe Coutinho	52	18	7	3.566245	3.433755
10	1633	West Ham United	7941	J. Hernández	28	16	8	4.733130	3.266870
11	1644	Watford	28292	A. Doucouré	44	17	7	3.770038	3.229962
12	1651	Brighton & Hove Albion	15526	P. Groß	41	19	7	3.878759	3.121241
13	1673	Huddersfield Town	38377	L. Depoitre	20	10	6	2.942375	3.057625
14	1646	Burnley	9206	C. Wood	38	25	10	7.040081	2.959919
15	1609	Arsenal	25867	P. Aubameyang	29	18	10	7.094684	2.905316
16	1625	Manchester City	11066	R. Sterling	80	35	18	15.184169	2.815831
17	1609	Arsenal	3560	Nacho Monreal	21	8	5	2.201240	2.798760
18	1659	AFC Bournemouth	62224	R. Fraser	31	13	5	2.216259	2.783741
19	1625	Manchester City	265673	Bernardo Silva	23	14	6	3.228071	2.771929

(4) 선수별 득점당 평균 xG 집계 및 정렬¶

In [9]:

player_stats['xG for goals'] = goals.groupby(['team_id', 'team_name', 'player_id', 'player_name'])['xg'].sum()
player_stats['xG per goal'] = player_stats['xG for goals'] / player_stats['Goals']
player_stats[player_stats['Goals'] >= 5].sort_values('xG per goal')[:20].reset_index()

Out[9]:

	team_id	team_name	player_id	player_name	Shots	SoT	Goals	xG	dG	xG for goals	xG per goal
0	1625	Manchester City	38021	K. De Bruyne	91	39	7	5.650185	1.349815	0.463997	0.066285
1	1625	Manchester City	105339	Fernandinho	48	20	5	3.116935	1.883065	0.421568	0.084314
2	1639	Stoke City	49872	X. Shaqiri	67	29	8	5.382163	2.617837	0.741266	0.092658
3	1631	Leicester City	26150	R. Mahrez	71	36	12	5.961230	6.038770	1.184440	0.098703
4	1659	AFC Bournemouth	62224	R. Fraser	31	13	5	2.216259	2.783741	0.535540	0.107108
5	1651	Brighton & Hove Albion	91381	J. Izquierdo	53	16	5	4.706987	0.293013	0.607775	0.121555
6	1644	Watford	28292	A. Doucouré	44	17	7	3.770038	3.229962	0.941342	0.134477
7	1612	Liverpool	3802	Philippe Coutinho	52	18	7	3.566245	3.433755	1.005339	0.143620
8	1611	Manchester United	397178	M. Rashford	59	21	7	5.264043	1.735957	1.008644	0.144092
9	1673	Huddersfield Town	38377	L. Depoitre	20	10	6	2.942375	3.057625	0.874010	0.145668
10	1644	Watford	20593	R. Pereyra	34	13	5	4.443552	0.556448	0.736922	0.147384
11	1625	Manchester City	245364	L. Sané	54	21	10	5.104557	4.895443	1.508979	0.150898
12	1624	Tottenham Hotspur	54	C. Eriksen	97	40	11	8.570227	2.429773	1.662022	0.151093
13	1610	Chelsea	8032	Marcos Alonso	59	21	7	5.528816	1.471184	1.082327	0.154618
14	1624	Tottenham Hotspur	14911	Son Heung-Min	70	31	12	6.654574	5.345426	1.922408	0.160201
15	1627	West Bromwich Albion	3577	S. Rondón	79	28	7	7.596607	-0.596607	1.127666	0.161095
16	1609	Arsenal	7870	A. Ramsey	52	22	7	6.215151	0.784849	1.142881	0.163269
17	1673	Huddersfield Town	214654	S. Mounié	49	21	8	6.587816	1.412184	1.307039	0.163380
18	1610	Chelsea	105333	Willian	53	25	6	5.184140	0.815860	1.023569	0.170595
19	1633	West Ham United	41174	M. Lanzini	43	16	5	4.712147	0.287853	0.854151	0.170830

선수별 시즌 전체 xG 시각화¶

(1) 슈팅 레이블 생성¶

In [10]:

match_df = pd.read_csv('data/refined_events/England/matches.csv', header=0, encoding='utf-8-sig')
shots = pd.merge(shots, match_df[['match_id', 'gameweek', 'team1_name', 'team2_name']])

# mouse over 시 보일 텍스트
shots['display_name'] = shots.apply(
    lambda x: f"[{x['gameweek']}R] {x['team1_name']} vs {x['team2_name']}, " +
    f"{x['period']} {int(x['time'] // 60):02d}:{int(x['time'] % 60):02d}, xG: {round(x['xg'], 3)}", axis=1
)
shots

Out[10]:

	competition_name	match_id	event_id	period	time	team_id	team_name	player_id	player_name	event_type	...	distance	angle	freekick	header	goal	xg	gameweek	team1_name	team2_name	display_name
0	England	2499719	177959212	1H	94.596	1609	Arsenal	25413	A. Lacazette	Shot	...	13.899813	26.933236	0	0	1	0.143848	1	Arsenal	Leicester City	[1R] Arsenal vs Leicester City, 1H 01:34, xG: ...
1	England	2499719	177959247	1H	179.855	1631	Leicester City	26150	R. Mahrez	Shot	...	15.659170	26.224941	0	0	0	0.120612	1	Arsenal	Leicester City	[1R] Arsenal vs Leicester City, 1H 02:59, xG: ...
2	England	2499719	177959280	1H	254.745	1631	Leicester City	14763	S. Okazaki	Shot	...	4.376665	79.289489	0	1	1	0.412967	1	Arsenal	Leicester City	[1R] Arsenal vs Leicester City, 1H 04:14, xG: ...
3	England	2499719	177959289	1H	425.824	1609	Arsenal	7868	A. Oxlade-Chamberlain	Shot	...	22.893038	15.813597	0	0	0	0.042434	1	Arsenal	Leicester City	[1R] Arsenal vs Leicester City, 1H 07:05, xG: ...
4	England	2499719	177959429	1H	815.462	1609	Arsenal	7868	A. Oxlade-Chamberlain	Shot	...	29.342120	12.655803	0	0	0	0.019414	1	Arsenal	Leicester City	[1R] Arsenal vs Leicester City, 1H 13:35, xG: ...
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
8876	England	2500089	251701666	2H	1682.852	1659	AFC Bournemouth	9637	J. King	Shot	...	15.367134	15.291389	0	0	1	0.083744	38	Burnley	AFC Bournemouth	[38R] Burnley vs AFC Bournemouth, 2H 28:02, xG...
8877	England	2500089	251701767	2H	2267.378	1659	AFC Bournemouth	9739	J. Ibe	Shot	...	26.432132	15.521739	0	0	0	0.029310	38	Burnley	AFC Bournemouth	[38R] Burnley vs AFC Bournemouth, 2H 37:47, xG...
8878	England	2500089	251701808	2H	2593.305	1659	AFC Bournemouth	11669	C. Wilson	Shot	...	32.824990	12.113593	0	0	0	0.013519	38	Burnley	AFC Bournemouth	[38R] Burnley vs AFC Bournemouth, 2H 43:13, xG...
8879	England	2500089	251701582	2H	2644.663	1646	Burnley	9179	N. Wells	Shot	...	18.176028	18.584769	0	0	0	0.073772	38	Burnley	AFC Bournemouth	[38R] Burnley vs AFC Bournemouth, 2H 44:04, xG...
8880	England	2500089	251701853	2H	2839.356	1659	AFC Bournemouth	11669	C. Wilson	Shot	...	11.736848	31.473639	0	0	1	0.188783	38	Burnley	AFC Bournemouth	[38R] Burnley vs AFC Bournemouth, 2H 47:19, xG...

8881 rows × 24 columns

(2) 슈팅 위치 및 xG 시각화¶

In [11]:

# 손흥민 선수의 xG 시각화
player_name = 'Son Heung-Min'
# player_name = 'Mohamed Salah'
# player_name = 'K. De Bruyne'

player_shots = shots[shots['player_name'] == player_name]
player_goals = player_shots[player_shots['tags'].apply(lambda x: 'Goal' in x)]
player_shots_failed = player_shots[player_shots['tags'].apply(lambda x: 'Goal' not in x)]

goal_trace = go.Scatter(
    x=104 - player_goals['x'],
    y=34 + player_goals['y'],
    name='Goal',
    text=player_goals['display_name'],
    mode='markers',
    marker=dict(
        color='red', size=np.sqrt(player_goals['xg']) * 50, 
        symbol=player_goals['freekick'].apply(lambda x: 'square' if x == 1 else 'circle')
    )
)

shot_trace = go.Scatter(
    x=104 - player_shots_failed['x'],
    y=34 + player_shots_failed['y'],
    name='Failed shot',
    text=player_shots_failed['display_name'],
    mode='markers',
    marker=dict(
        color='darkgrey', size=np.sqrt(player_shots_failed['xg']) * 50, 
        symbol=player_shots_failed['freekick'].apply(lambda x: 'square' if x == 1 else 'circle')
    )
)

title = f"{player_name} - Goals: {player_shots['goal'].sum()}, xG: {round(player_shots['xg'].sum(), 3)} "
fig = go.Figure(data=[shot_trace, goal_trace], layout=get_pitch_layout(title))
fig.show()