[Football Analysis] 06. 시간 구간별 이벤트 집계 및 시각화

작성일자 : 2023-11-11

Ver 0.1.1

 

In [1]:

import os
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt

In [2]:

# 현대 디렉토리 확인
current_dir = os.getcwd()
current_dir

Out[2]:

'/Users/limjongjun/Desktop/JayJay/Growth/Python/soccer-analytics/Excercise'

In [3]:

# Data가 있는 디렉토리로 변경
new_dir = '/Users/limjongjun/Desktop/JayJay/Growth/Python/soccer-analytics'
os.chdir(new_dir)
print(current_dir)

/Users/limjongjun/Desktop/JayJay/Growth/Python/soccer-analytics/Excercise

In [4]:

# Manshester City VS Tottenham Hotspur 경기 불러오기
match_id = 2499895
match_events = pd.read_pickle(f'data/refined_events/England/2499895.pkl') #pkl 파일을 Dataframe으로 불러오기

In [5]:

# 데이터 확인
match_events.head()

Out[5]:

	match_id	event_id	period	time	team_id	team_name	player_id	player_name	event_type	sub_event_type	tags	start_x	start_y	end_x	end_y
0	2499895	215108367	1H	1.784	1625	Manchester City	8325	S. Agüero	Pass	Simple pass	[Accurate]	52.00	34.68	40.56	34.68
1	2499895	215108368	1H	3.324	1625	Manchester City	105339	Fernandinho	Pass	Simple pass	[Accurate]	40.56	34.68	29.12	8.16
2	2499895	215108369	1H	6.406	1625	Manchester City	8277	K. Walker	Pass	Simple pass	[Accurate]	29.12	8.16	44.72	10.20
3	2499895	215108370	1H	7.124	1625	Manchester City	38021	K. De Bruyne	Pass	Simple pass	[Accurate]	44.72	10.20	69.68	6.12
4	2499895	215108371	1H	8.676	1625	Manchester City	11066	R. Sterling	Pass	Simple pass	[Accurate]	69.68	6.12	58.24	12.24

경기 이벤트 시간 구간별 집계 및 시각화¶

(1) pandas.cut 함수를 활용한 연속형 데이터 분할¶

• 5분 단위 분할 집계를 위한 기준점 리스트 생성

In [6]:

bins = np.arange(0,50,5)
bins = np.append(bins,[100])
bins

Out[6]:

array([  0,   5,  10,  15,  20,  25,  30,  35,  40,  45, 100])

• 기준점을 이용한 시간 구간 분할

In [7]:

match_events['time_cat'] = pd.cut(match_events['time'] / 60 , bins, right = False)

In [8]:

match_events

Out[8]:

	match_id	event_id	period	time	team_id	team_name	player_id	player_name	event_type	sub_event_type	tags	start_x	start_y	end_x	end_y	time_cat
0	2499895	215108367	1H	1.784	1625	Manchester City	8325	S. Agüero	Pass	Simple pass	[Accurate]	52.00	34.68	40.56	34.68	[0, 5)
1	2499895	215108368	1H	3.324	1625	Manchester City	105339	Fernandinho	Pass	Simple pass	[Accurate]	40.56	34.68	29.12	8.16	[0, 5)
2	2499895	215108369	1H	6.406	1625	Manchester City	8277	K. Walker	Pass	Simple pass	[Accurate]	29.12	8.16	44.72	10.20	[0, 5)
3	2499895	215108370	1H	7.124	1625	Manchester City	38021	K. De Bruyne	Pass	Simple pass	[Accurate]	44.72	10.20	69.68	6.12	[0, 5)
4	2499895	215108371	1H	8.676	1625	Manchester City	11066	R. Sterling	Pass	Simple pass	[Accurate]	69.68	6.12	58.24	12.24	[0, 5)
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
1607	2499895	215110122	2H	2875.703	1624	Tottenham Hotspur	210044	E. Dier	Pass	Simple pass	[Accurate]	36.40	26.52	48.88	29.92	[45, 100)
1608	2499895	215110123	2H	2876.142	1624	Tottenham Hotspur	240070	H. Winks	Duel	Ground attacking duel	[Anticipation, Lost, Not accurate]	48.88	29.92	43.68	24.48	[45, 100)
1609	2499895	215109959	2H	2876.768	1625	Manchester City	447205	P. Foden	Duel	Ground defending duel	[Anticipated, Won, Accurate]	55.12	38.08	60.32	43.52	[45, 100)
1610	2499895	215109960	2H	2878.046	1625	Manchester City	11066	R. Sterling	Duel	Ground attacking duel	[Free space right, Lost, Not accurate]	60.32	43.52	60.32	43.52	[45, 100)
1611	2499895	215110124	2H	2878.216	1624	Tottenham Hotspur	240070	H. Winks	Duel	Ground defending duel	[Free space left, Lost, Not accurate]	43.68	24.48	NaN	NaN	[45, 100)

1612 rows × 16 columns

(2) 시간 구간별 이벤트 횟수 집계¶

• 5분 단위 팀당 이벤트 횟수 집계

In [9]:

# 주어진 데이터프레임(match_events)에서 'event_id' 열을 사용하여 피벗 테이블을 생성합니다.
# 각 피벗 테이블 셀은 'event_id' 값들에 대한 통계 정보를 담게 됩니다.
# 인덱스로는 'period' 및 'time_cat' 열의 고유한 조합을 사용하며,
# 각 팀('team_name')은 피벗 테이블의 열로 사용됩니다.
# 'count' 함수는 'event_id' 값들의 개수를 세기 위해 사용됩니다.
match_events.pivot_table(values='event_id',  # 'event_id' 열의 값을 사용
                         index=['period', 'time_cat'],  # 'period' 및 'time_cat' 열을 인덱스로 사용
                         columns='team_name',  # 'team_name' 열을 피벗 테이블의 열로 사용
                         aggfunc='count')  # 'count' 함수를 사용하여 값들을 집계

# match_events.pivot_table('event_id',  ['period','time_cat'] , 'team_name', 'count')

Out[9]:

	team_name	Manchester City	Tottenham Hotspur
period	time_cat
1H	[0, 5)	56	56
	[5, 10)	62	48
	[10, 15)	51	38
	[15, 20)	43	53
	[20, 25)	28	55
	[25, 30)	52	33
	[30, 35)	28	42
	[35, 40)	20	35
	[40, 45)	60	35
	[45, 100)	22	20
2H	[0, 5)	48	57
	[5, 10)	29	59
	[10, 15)	37	37
	[15, 20)	43	46
	[20, 25)	40	30
	[25, 30)	39	46
	[30, 35)	41	47
	[35, 40)	14	35
	[40, 45)	56	27
	[45, 100)	23	21

• 시간 구간 레이블 생성

In [10]:

labels = []
for i in range(len(bins)-2):
    label = f'{bins[i]:02d}-{bins[i+1]:02d}'
    labels.append(label)
labels.append('45+')
labels

Out[10]:

['00-05',
 '05-10',
 '10-15',
 '15-20',
 '20-25',
 '25-30',
 '30-35',
 '35-40',
 '40-45',
 '45+']

• 시간 구간 레이블을 추가한 5분 단위 팀당 이벤트 횟수 집계

In [11]:

match_events['time_cat'] = pd.cut(match_events['time'] / 60 , bins = bins , labels=labels, right = False)
count_by_time = match_events.pivot_table( 'event_id',  ['period','time_cat'] , 'team_name', 'count', fill_value=0)
count_by_time

Out[11]:

	team_name	Manchester City	Tottenham Hotspur
period	time_cat
1H	00-05	56	56
	05-10	62	48
	10-15	51	38
	15-20	43	53
	20-25	28	55
	25-30	52	33
	30-35	28	42
	35-40	20	35
	40-45	60	35
	45+	22	20
2H	00-05	48	57
	05-10	29	59
	10-15	37	37
	15-20	43	46
	20-25	40	30
	25-30	39	46
	30-35	41	47
	35-40	14	35
	40-45	56	27
	45+	23	21

(3) 전반전 이벤트 횟수 5분 단위 시각화¶

• 전반전 시간 구간별 전체 이벤트 횟수 집계

In [12]:

half_count_by_time = count_by_time.loc['1H', 'Manchester City']
half_count_by_time

Out[12]:

time_cat
00-05    56
05-10    62
10-15    51
15-20    43
20-25    28
25-30    52
30-35    28
35-40    20
40-45    60
45+      22
Name: Manchester City, dtype: int64

• plt.bar 함수를 활용한 막대그래프 시각화

In [13]:

plt.bar(half_count_by_time.index, half_count_by_time.values)
plt.show()

(4) 경기 전체 이벤트 횟수 5분 단위 시각화¶

• 5분 단위 팀당 이벤트 횟수

In [14]:

count_by_time

Out[14]:

	team_name	Manchester City	Tottenham Hotspur
period	time_cat
1H	00-05	56	56
	05-10	62	48
	10-15	51	38
	15-20	43	53
	20-25	28	55
	25-30	52	33
	30-35	28	42
	35-40	20	35
	40-45	60	35
	45+	22	20
2H	00-05	48	57
	05-10	29	59
	10-15	37	37
	15-20	43	46
	20-25	40	30
	25-30	39	46
	30-35	41	47
	35-40	14	35
	40-45	56	27
	45+	23	21

• 다중 인덱스를 단일 인덱스로 변환

In [15]:

idx_1h = []
idx_2h = []

for multi_idx in count_by_time.index:
    if multi_idx[0] == '1H':
        idx = multi_idx[1]
        idx_1h.append(idx)
    else:
        if '-' in multi_idx[1]:
            edges = multi_idx[1].split('-')
            start_min = int(edges[0]) + 45
            end_min = int(edges[1]) + 45
            idx = f'{start_min:02d}-{end_min:02d}'
        else:
            idx = '90+'
        idx_2h.append(idx)

idx_1h, idx_2h

Out[15]:

(['00-05',
  '05-10',
  '10-15',
  '15-20',
  '20-25',
  '25-30',
  '30-35',
  '35-40',
  '40-45',
  '45+'],
 ['45-50',
  '50-55',
  '55-60',
  '60-65',
  '65-70',
  '70-75',
  '75-80',
  '80-85',
  '85-90',
  '90+'])

• plt.barh 함수를 활용한 수평 막대 그래프 시각화

In [16]:

count_by_time.index = idx_1h + idx_2h
count_by_time

Out[16]:

team_name	Manchester City	Tottenham Hotspur
00-05	56	56
05-10	62	48
10-15	51	38
15-20	43	53
20-25	28	55
25-30	52	33
30-35	28	42
35-40	20	35
40-45	60	35
45+	22	20
45-50	48	57
50-55	29	59
55-60	37	37
60-65	43	46
65-70	40	30
70-75	39	46
75-80	41	47
80-85	14	35
85-90	56	27
90+	23	21

In [17]:

plt.figure(figsize=(6,6))

for half, half_idx, color in zip(['1H','2H'], [idx_1h, idx_2h], ['skyblue', 'blue']): #전반은 하늘색, 후반전은 파란색
    half_counts = count_by_time['Manchester City'].loc[half_idx]
    plt.barh(half_idx, half_counts, label = half, color = color, edgecolor = 'black') #수직 막대그래프 

plt.gca().invert_yaxis() #y축 반전
plt.grid(axis= 'x', color = 'k', linestyle = '--')
plt.legend(fontsize= 15)
plt.xticks(fontsize= 15)
plt.yticks(fontsize= 15)
plt.xlabel('Number of events', fontsize= 20)
plt.ylabel('Match time (min)', fontsize= 20)
plt.show()

In [18]:

fig, axes = plt.subplots(ncols = 2, sharey = True, figsize = (12,6)) #sharey = True : y 축 공유
team1_name, team2_name = match_events['team_name'].unique()

for half, half_idx, color in zip(['1H','2H'], [idx_1h, idx_2h], ['skyblue', 'blue']): #전반은 하늘색, 후반전은 파란색
    half_counts = count_by_time[team1_name].loc[half_idx] # Manchester City
    axes[0].barh(half_idx, half_counts, label = half, color = color, edgecolor = 'k')

for half, half_idx, color in zip(['1H','2H'], [idx_1h, idx_2h], ['grey', 'black']): #전반은 회색, 후반전은 검정
    half_counts = count_by_time[team2_name].loc[half_idx] # Tottenham
    axes[1].barh(half_idx, half_counts, label = half, color = color, edgecolor = 'k')

plt.gca().invert_yaxis()

axes[0].invert_xaxis()
axes[0].yaxis.tick_right()

for team_name, ax in zip([team1_name, team2_name], axes):
    ax.grid(axis = 'x', color = 'k', linestyle = '--')
    ax.set_title(f'#. Event by {team_name}', size = 18, fontweight = 'bold')
    ax.set_xticks(np.arange(0,100,20))
    ax.tick_params(labelsize = 15)

plt.show()

대회 전체 이벤트 데이터 시간 구간별 집계 및 시각화¶

(1) 대회 전체 이벤트 데이터를 하나의 DataFrame으로 연결하기¶

In [19]:

dataset_name = 'England'
match_df = pd.read_csv(f'data/refined_events/England/matches.csv', index_col=0, encoding='utf-8-sig')

match_events_list = []

for match_id in match_df.index :
    match_events = pd.read_pickle(f'data/refined_events/England/{match_id}.pkl')
    match_events_list.append(match_events)

events = pd.concat(match_events_list, ignore_index= True)
events

Out[19]:

	match_id	event_id	period	time	team_id	team_name	player_id	player_name	event_type	sub_event_type	tags	start_x	start_y	end_x	end_y
0	2499719	177959171	1H	2.759	1609	Arsenal	25413	A. Lacazette	Pass	Simple pass	[Accurate]	50.96	34.68	32.24	14.96
1	2499719	177959172	1H	4.947	1609	Arsenal	370224	R. Holding	Pass	High pass	[Accurate]	32.24	14.96	53.04	17.00
2	2499719	177959173	1H	6.542	1609	Arsenal	3319	M. Özil	Pass	Head pass	[Accurate]	53.04	17.00	36.40	19.72
3	2499719	177959174	1H	8.143	1609	Arsenal	120339	Mohamed Elneny	Pass	Head pass	[Accurate]	36.40	19.72	42.64	3.40
4	2499719	177959175	1H	10.302	1609	Arsenal	167145	Bellerín	Pass	Simple pass	[Accurate]	42.64	3.40	74.88	8.16
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
647247	2500089	251701859	2H	3010.063	1659	AFC Bournemouth	11669	C. Wilson	Pass	Head pass	[Not accurate]	63.44	47.60	65.52	31.96
647248	2500089	251701668	2H	3012.989	1646	Burnley	8433	S. Ward	Pass	High pass	[Not accurate]	38.48	36.04	74.88	6.80
647249	2500089	251701860	2H	3014.732	1659	AFC Bournemouth	134102	N. Aké	Others on the ball	Clearance	[Interception, Not accurate]	29.12	61.20	28.08	68.00
647250	2500089	251701861	2H	3018.672	1659	AFC Bournemouth	0	NaN	Interruption	Ball out of the field	[]	28.08	68.00	NaN	NaN
647251	2500089	251701670	2H	3021.271	1646	Burnley	8980	J. Hendrick	Free kick	Throw in	[Not accurate]	70.72	0.00	NaN	NaN

647252 rows × 15 columns

In [20]:

bins = np.arange(0,50,5)
bins = np.append(bins,[100])

labels = []
for i in range(len(bins)-2):
    label = f'{bins[i]:02d}-{bins[i+1]:02d}'
    labels.append(label)
labels.append('45+')

events['time_cat'] = pd.cut(events['time'] / 60 , bins = bins , labels=labels, right = False)
count_by_time = events.groupby(['period', 'time_cat'])['event_id'].count()
count_by_time

Out[20]:

period  time_cat
1H      00-05       37582
        05-10       35909
        10-15       35278
        15-20       34003
        20-25       33973
        25-30       33925
        30-35       33390
        35-40       33357
        40-45       33717
        45+         12610
2H      00-05       36706
        05-10       34224
        10-15       32706
        15-20       32440
        20-25       32347
        25-30       32077
        30-35       31971
        35-40       32499
        40-45       32063
        45+         26475
Name: event_id, dtype: int64

In [21]:

mean_count_by_time = count_by_time.loc[['1H','2H']] / len(match_df)

idx_1h = []
idx_2h = []

for multi_idx in mean_count_by_time.index:
    if multi_idx[0] == '1H':
        idx = multi_idx[1]
        idx_1h.append(idx)
    else:
        if '-' in multi_idx[1]:
            edges = multi_idx[1].split('-')
            start_min = int(edges[0]) + 45
            end_min = int(edges[1]) + 45
            idx = f'{start_min:02d}-{end_min:02d}'
        else:
            idx = '90+'
        idx_2h.append(idx)

mean_count_by_time.index = idx_1h + idx_2h
mean_count_by_time

Out[21]:

00-05    98.900000
05-10    94.497368
10-15    92.836842
15-20    89.481579
20-25    89.402632
25-30    89.276316
30-35    87.868421
35-40    87.781579
40-45    88.728947
45+      33.184211
45-50    96.594737
50-55    90.063158
55-60    86.068421
60-65    85.368421
65-70    85.123684
70-75    84.413158
75-80    84.134211
80-85    85.523684
85-90    84.376316
90+      69.671053
Name: event_id, dtype: float64

In [24]:

plt.figure(figsize=(16,6))

for half, half_idx, color in zip(['1H','2H'], [idx_1h, idx_2h], ['C0', 'C1']):
    half_counts = mean_count_by_time[half_idx]
    plt.bar(half_idx, half_counts, label = half, color = color) # 수평 막대 그래프

plt.grid(axis= 'y', color = 'k', linestyle = '--')
plt.legend(fontsize= 15)
plt.xticks(fontsize= 15)
plt.yticks(fontsize= 15)
plt.xlabel('Match time (min)', fontsize= 20)
plt.ylabel('Number of events', fontsize= 20)
plt.show()

(3) 시간 구간별 득점 횟수 집계 및 시각화¶

• 득점 데이터 추출

In [25]:

goals = events[(events['tags'].apply(lambda x: 'Goal' in x or 'Own goal' in x))]
goals

Out[25]:

	match_id	event_id	period	time	team_id	team_name	player_id	player_name	event_type	sub_event_type	tags	start_x	start_y	end_x	end_y	time_cat
46	2499719	177959212	1H	94.596	1609	Arsenal	25413	A. Lacazette	Shot	Shot	[Goal, Right foot, Opportunity, Position: Goal...	91.52	40.12	104.0	34.0	00-05
91	2499719	177959280	1H	254.745	1631	Leicester City	14763	S. Okazaki	Shot	Shot	[Goal, Head/body, Opportunity, Position: Goal ...	99.84	32.64	104.0	34.0	00-05
554	2499719	177959759	1H	1710.855	1631	Leicester City	12829	J. Vardy	Shot	Shot	[Goal, Right foot, Opportunity, Position: Goal...	97.76	31.28	104.0	34.0	25-30
898	2499719	177960130	1H	2815.901	1609	Arsenal	7945	D. Welbeck	Shot	Shot	[Goal, Right foot, Opportunity, Position: Goal...	95.68	34.00	104.0	34.0	45+
1107	2499719	177960379	2H	634.312	1631	Leicester City	12829	J. Vardy	Shot	Shot	[Goal, Head/body, Opportunity, Position: Goal ...	95.68	31.28	104.0	34.0	10-15
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
645400	2500098	251596263	2H	1737.065	1623	Everton	145692	O. Niasse	Shot	Shot	[Goal, Right foot, Opportunity, Position: Goal...	100.88	30.60	104.0	34.0	25-30
645535	2500098	251596096	2H	2193.887	1633	West Ham United	41174	M. Lanzini	Shot	Shot	[Goal, Right foot, Opportunity, Position: Goal...	89.44	45.56	104.0	34.0	35-40
646362	2500089	251700843	1H	2307.084	1646	Burnley	9206	C. Wood	Shot	Shot	[Goal, Left foot, Interception, Opportunity, P...	95.68	34.68	104.0	34.0	35-40
646916	2500089	251701666	2H	1682.852	1659	AFC Bournemouth	9637	J. King	Shot	Shot	[Goal, Right foot, Opportunity, Position: Goal...	95.68	46.92	104.0	34.0	25-30
647226	2500089	251701853	2H	2839.356	1659	AFC Bournemouth	11669	C. Wilson	Shot	Shot	[Goal, Right foot, Opportunity, Position: Goal...	93.60	28.56	104.0	34.0	45+

1017 rows × 16 columns

In [26]:

goals_by_time = goals.groupby(['period', 'time_cat'])['event_id'].count()[['1H', '2H']]
goals_by_time.index = idx_1h + idx_2h
goals_by_time

Out[26]:

00-05    37
05-10    40
10-15    42
15-20    50
20-25    48
25-30    49
30-35    41
35-40    47
40-45    47
45+      26
45-50    56
50-55    63
55-60    54
60-65    62
65-70    60
70-75    66
75-80    58
80-85    59
85-90    63
90+      49
Name: event_id, dtype: int64

In [27]:

plt.figure(figsize=(12, 6))

for half, half_idx, color in zip(['1H', '2H'], [idx_1h, idx_2h], ['C0', 'C1']):
    half_counts = goals_by_time[half_idx]
    plt.bar(half_idx, half_counts, label=half, color=color)

plt.grid(axis='y', color='k', linestyle='--')
plt.legend(fontsize=15)
plt.xticks(fontsize=15, rotation=90)
plt.yticks(np.arange(0, 70, 5), fontsize=15)
plt.xlabel('Match time (min)', fontsize=20)
plt.ylabel('Number of goals', fontsize=20)
plt.show()