NBA20-21赛季数据分析 | Python数据分析及可视化实战

数据的一般处理

import pandas as pd
import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt
import warnings
warnings.filterwarnings("ignore")
plt.rcParams['axes.unicode_minus'] = False
sns.set_style('darkgrid', {'font.sans-serif':['SimHei', 'Arial']})

df_player = pd.read_csv(r"C:\Users\imqqdong\Downloads\NBAdata\season_2021_detailed.csv")
df_team = pd.read_csv(r"C:\Users\imqqdong\Downloads\NBAdata\season_2021_basic.csv")

df_team.head()

	date	weekday	home_team	home_score	away_team	away_score	attendance	overtime	remarks
0	2020-12-22T19:00:00	Tuesday	Brooklyn Nets	125	Golden State Warriors	99	0	NaN	NaN
1	2020-12-22T22:00:00	Tuesday	Los Angeles Lakers	109	Los Angeles Clippers	116	0	NaN	NaN
2	2020-12-23T19:00:00	Wednesday	Cleveland Cavaliers	121	Charlotte Hornets	114	0	NaN	NaN
3	2020-12-23T19:00:00	Wednesday	Indiana Pacers	121	New York Knicks	107	0	NaN	NaN
4	2020-12-23T19:00:00	Wednesday	Orlando Magic	113	Miami Heat	107	0	NaN	NaN

df_team_sim = df_team.drop(columns=["weekday","attendance","overtime","remarks"])

df_team_sim["winner"] = 0
df_team_sim["loser"] = 0
for i in range(len(df_team_sim)):
    if df_team_sim["home_score"][i] > df_team_sim["away_score"][i]:
        df_team_sim["winner"][i] = df_team_sim["home_team"][i]
        df_team_sim["loser"][i] = df_team_sim["away_team"][i]
    elif df_team_sim["home_score"][i] < df_team_sim["away_score"][i]:
        df_team_sim["winner"][i] = df_team_sim["away_team"][i]
        df_team_sim["loser"][i] = df_team_sim["home_team"][i]

sns.countplot(x = df_team_sim["winner"],color = "orange")
plt.xticks(rotation = 90)
plt.show() 

win_lose_times = pd.DataFrame(df_team_sim["winner"].value_counts())

win_lose_times["lose_times"] = df_team_sim["loser"].value_counts()
#会不会自动对齐索引

win_lose_times.head()

	winner	lose_times
Phoenix Suns	63	25
Utah Jazz	58	25
Los Angeles Clippers	57	34
Milwaukee Bucks	57	31
Philadelphia 76ers	56	28

win_lose_times.rename(columns={"winner":"win_times"},inplace = True)

fig= plt.figure(figsize = (20,10),dpi = 100)
win_lose_times.plot(y = ["win_times","lose_times"],kind = "bar"
                    ,use_index = True,title = "2020-2021赛季NBA各队伍胜负场次",
                   grid = True,)

<AxesSubplot:title={'center':'2020-2021赛季NBA各队伍胜负场次'}>




<Figure size 2000x1000 with 0 Axes>

本赛季各个队伍的胜率情况

win_lose_times["wp"] = win_lose_times["win_times"]/(win_lose_times["win_times"]+win_lose_times["lose_times"])

win_lose_times.sort_values(by = "wp",ascending=False)

	win_times	lose_times	wp
Phoenix Suns	63	25	0.715909
Utah Jazz	58	25	0.698795
Philadelphia 76ers	56	28	0.666667
Brooklyn Nets	55	29	0.654762
Milwaukee Bucks	57	31	0.647727
Los Angeles Clippers	57	34	0.626374
Denver Nuggets	51	31	0.621951
Atlanta Hawks	51	38	0.573034
Dallas Mavericks	45	34	0.569620
Los Angeles Lakers	45	34	0.569620
Portland Trail Blazers	44	34	0.564103
New York Knicks	42	35	0.545455
Golden State Warriors	39	35	0.527027
Miami Heat	40	36	0.526316
Memphis Grizzlies	41	38	0.518987
Boston Celtics	38	40	0.487179
Indiana Pacers	35	39	0.472973
Washington Wizards	36	43	0.455696
San Antonio Spurs	33	40	0.452055
Charlotte Hornets	33	40	0.452055
New Orleans Pelicans	31	41	0.430556
Chicago Bulls	31	41	0.430556
Sacramento Kings	31	41	0.430556
Toronto Raptors	27	45	0.375000
Minnesota Timberwolves	23	49	0.319444
Oklahoma City Thunder	22	50	0.305556
Cleveland Cavaliers	22	50	0.305556
Orlando Magic	21	51	0.291667
Detroit Pistons	20	52	0.277778
Houston Rockets	17	55	0.236111

fig = plt.figure(figsize = (30,20),dpi =100)
x = np.arange(len(win_lose_times.index.values))
width = 0.4
plt.bar(x = x - width/2,height = win_lose_times["win_times"],label = "胜利场次",align="center",width = width)
plt.bar(x = x + width/2,height = win_lose_times["lose_times"],label = "失败场次",align="center",width = width)
plt.xticks(ticks = x,rotation = 90,fontsize =25,labels = win_lose_times.index.values)
plt.yticks(fontsize = 25)
plt.legend(fontsize = 25)
plt.show()

df_player.head()

	date	team	player	role	MP	FG	FGA	FG_PCT	FG3	FG3A	...	ORB	DRB	TRB	AST	STL	BLK	TOV	PF	PTS	PLUS_MINUS
0	2020-12-22T19:00:00	Brooklyn Nets	Tyler Johnson	Reserve	5.083333	0	1	0.000	0	0	...	1	0	1	0	1	0	0	0	0	-12.0
1	2020-12-22T19:00:00	Brooklyn Nets	Kyrie Irving	Starter	25.416667	10	16	0.625	4	7	...	1	3	4	4	0	0	1	3	26	32.0
2	2020-12-22T19:00:00	Brooklyn Nets	Kevin Durant	Starter	24.400000	7	16	0.438	1	2	...	1	4	5	3	3	1	1	3	22	26.0
3	2020-12-22T19:00:00	Brooklyn Nets	Joe Harris	Starter	20.333333	4	8	0.500	2	5	...	2	5	7	2	0	1	2	2	10	21.0
4	2020-12-22T19:00:00	Brooklyn Nets	Spencer Dinwiddie	Starter	19.316667	2	6	0.333	1	3	...	0	5	5	4	0	0	3	1	9	24.0

5 rows × 24 columns

#看看詹姆斯这个赛季的数据
df_james = df_player[df_player["player"]=="LeBron James"]

df_james.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 52 entries, 50 to 24083
Data columns (total 24 columns):
 #   Column      Non-Null Count  Dtype  
---  ------      --------------  -----  
 0   date        52 non-null     object 
 1   team        52 non-null     object 
 2   player      52 non-null     object 
 3   role        52 non-null     object 
 4   MP          52 non-null     float64
 5   FG          52 non-null     int64  
 6   FGA         52 non-null     int64  
 7   FG_PCT      52 non-null     float64
 8   FG3         52 non-null     int64  
 9   FG3A        52 non-null     int64  
 10  FG3_PCT     52 non-null     float64
 11  FT          52 non-null     int64  
 12  FTA         52 non-null     int64  
 13  FT_PCT      51 non-null     float64
 14  ORB         52 non-null     int64  
 15  DRB         52 non-null     int64  
 16  TRB         52 non-null     int64  
 17  AST         52 non-null     int64  
 18  STL         52 non-null     int64  
 19  BLK         52 non-null     int64  
 20  TOV         52 non-null     int64  
 21  PF          52 non-null     int64  
 22  PTS         52 non-null     int64  
 23  PLUS_MINUS  52 non-null     float64
dtypes: float64(5), int64(15), object(4)
memory usage: 10.2+ KB

#处理一下空值
pd.set_option("display.max_columns",None)
df_james.loc[df_james["FT_PCT"].isnull() == True,:]

	date	team	player	role	MP	FG	FGA	FG_PCT	FG3	FG3A	FG3_PCT	FT	FTA	FT_PCT	ORB	DRB	TRB	AST	STL	BLK	TOV	PF	PTS	PLUS_MINUS
23939	2021-06-01T22:00:00	Los Angeles Lakers	LeBron James	Starter	31.516667	9	19	0.474	6	10	0.6	0	0	NaN	1	4	5	7	0	0	3	1	24	-24.0

df_james.fillna(value = 0,inplace = True)

df_james.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 52 entries, 50 to 24083
Data columns (total 24 columns):
 #   Column      Non-Null Count  Dtype  
---  ------      --------------  -----  
 0   date        52 non-null     object 
 1   team        52 non-null     object 
 2   player      52 non-null     object 
 3   role        52 non-null     object 
 4   MP          52 non-null     float64
 5   FG          52 non-null     int64  
 6   FGA         52 non-null     int64  
 7   FG_PCT      52 non-null     float64
 8   FG3         52 non-null     int64  
 9   FG3A        52 non-null     int64  
 10  FG3_PCT     52 non-null     float64
 11  FT          52 non-null     int64  
 12  FTA         52 non-null     int64  
 13  FT_PCT      52 non-null     float64
 14  ORB         52 non-null     int64  
 15  DRB         52 non-null     int64  
 16  TRB         52 non-null     int64  
 17  AST         52 non-null     int64  
 18  STL         52 non-null     int64  
 19  BLK         52 non-null     int64  
 20  TOV         52 non-null     int64  
 21  PF          52 non-null     int64  
 22  PTS         52 non-null     int64  
 23  PLUS_MINUS  52 non-null     float64
dtypes: float64(5), int64(15), object(4)
memory usage: 10.2+ KB

詹姆斯本赛季的平均出场时间

#詹姆斯本赛季场均出场时间33.94分钟
fig = plt.figure(figsize=(20,10),dpi= 100)
x = np.arange(len(df_james))
mp_mean = df_james["MP"].mean()
df_james["MP"].plot(kind = "bar")
plt.axhline(mp_mean)
plt.xticks(ticks=x,labels = df_james["date"])
plt.yticks(fontsize=15)
plt.text(x = 0,y = mp_mean+2,s = "Mean:%.2f"%mp_mean,fontsize = 20,)
plt.show()

现在看一下詹姆斯在场上的时候，湖人的胜率

#詹姆斯上场时，湖人的胜率
'''
这块的计算错了，也放在上面给大家看一下
主要原因是dataframe与字符串的比较
#df_james.date.values[i]
def LA_WP(dates):
'''
    '''
    。。。
    '''
    win_times=0
    lose_times=0
    for date in dates:
        #win_times = 0
        #lose_times = 0  
        winner = df_team_sim[df_team_sim["date"]==date].winner.values
        if winner.all() == "Los Angeles Lakers":
            win_times += 1
        else :
            lose_times+=1
    return ("胜利场次：%d,失败场次：%d,胜率：%.2f"%(win_times,lose_times,win_times/(win_times+lose_times)))

LA_WP(df_james.date.values)

'胜利场次：23,失败场次：29,胜率：0.44'

计算詹姆斯不在场时，湖人的胜率

先不管上面的错误，计算詹姆斯不在场时，湖人的胜率

#詹姆斯不在场时，湖人的胜率
#找出詹姆斯不在场的比赛的时间，用isin取反即可
df_nojames = df_team_sim[~df_team_sim["date"].isin(df_james["date"])]
df_nojames = df_nojames[(df_nojames["home_team"]=="Los Angeles Lakers" )|( df_nojames["away_team"]=="Los Angeles Lakers")]
len(df_nojames)

27

win_games = df_nojames[df_nojames["winner"] == "Los Angeles Lakers"]
lose_games = df_nojames[df_nojames["loser"]=="Los Angeles Lakers"]
wp_nojames = len(win_games)/(len(win_games)+len(lose_games))
wp_nojames

0.4444444444444444

lose_games

	date	home_team	home_score	away_team	away_score	winner	loser
523	2021-03-03T22:00:00	Sacramento Kings	123	Los Angeles Lakers	120	Sacramento Kings	Los Angeles Lakers
625	2021-03-21T22:00:00	Phoenix Suns	111	Los Angeles Lakers	94	Phoenix Suns	Los Angeles Lakers
635	2021-03-23T19:30:00	New Orleans Pelicans	128	Los Angeles Lakers	111	New Orleans Pelicans	Los Angeles Lakers
655	2021-03-25T22:00:00	Los Angeles Lakers	101	Philadelphia 76ers	109	Philadelphia 76ers	Los Angeles Lakers
704	2021-03-31T22:00:00	Los Angeles Lakers	97	Milwaukee Bucks	112	Milwaukee Bucks	Los Angeles Lakers
731	2021-04-04T15:30:00	Los Angeles Clippers	104	Los Angeles Lakers	86	Los Angeles Clippers	Los Angeles Lakers
761	2021-04-08T19:30:00	Miami Heat	110	Los Angeles Lakers	104	Miami Heat	Los Angeles Lakers
795	2021-04-12T19:30:00	New York Knicks	111	Los Angeles Lakers	96	New York Knicks	Los Angeles Lakers
824	2021-04-15T22:00:00	Los Angeles Lakers	113	Boston Celtics	121	Boston Celtics	Los Angeles Lakers
858	2021-04-19T22:00:00	Los Angeles Lakers	97	Utah Jazz	111	Utah Jazz	Los Angeles Lakers
881	2021-04-22T21:30:00	Dallas Mavericks	115	Los Angeles Lakers	110	Dallas Mavericks	Los Angeles Lakers
894	2021-04-24T20:30:00	Dallas Mavericks	108	Los Angeles Lakers	93	Dallas Mavericks	Los Angeles Lakers
923	2021-04-28T19:30:00	Washington Wizards	116	Los Angeles Lakers	107	Washington Wizards	Los Angeles Lakers
993	2021-05-06T22:00:00	Los Angeles Clippers	118	Los Angeles Lakers	94	Los Angeles Clippers	Los Angeles Lakers
1002	2021-05-07T22:00:00	Portland Trail Blazers	106	Los Angeles Lakers	101	Portland Trail Blazers	Los Angeles Lakers

#根据时间和队名的匹配，找出詹姆斯在场的每一个比赛，组合到一个dataframe中
data = pd.DataFrame()
for i in range(len(df_james)):
    james_game = df_team_sim[(df_team_sim["date"]==df_james["date"].values[i])
                               &((df_team_sim["winner"]=="Los Angeles Lakers")|(df_team_sim["loser"]=="Los Angeles Lakers"))]
    data = pd.concat([data,james_game])

#詹姆斯在场上的时候胜率wp_james
james_win = len(data[data["winner"]=="Los Angeles Lakers"])
james_lose = len(data[data["loser"]=="Los Angeles Lakers"])
wp_james = james_win/(james_win+james_lose)
wp_james

0.6346153846153846

不在场时胜率40%，在场时胜率60%，可见詹姆斯的对球队胜率的影响还是很大的。后续将继续关注其与戴维斯合作时的相关数据，将更注重数据的可视化展现…
后续更新…