NBA20-21赛季数据分析 | Python数据分析及可视化实战_nba球员赛季跑动距离数据可视化

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']})
1
2
3
4
5
6
7
8

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")
1
2

df_team.head()
1

df_team_sim = df_team.drop(columns=["weekday","attendance","overtime","remarks"])
1

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]

1
2
3
4
5
6
7
8
9
10

sns.countplot(x = df_team_sim["winner"],color = "orange")
plt.xticks(rotation = 90)
plt.show() 
1
2
3

win_lose_times = pd.DataFrame(df_team_sim["winner"].value_counts())
1

win_lose_times["lose_times"] = df_team_sim["loser"].value_counts()
#会不会自动对齐索引
1
2

win_lose_times.head()
1

win_lose_times.rename(columns={"winner":"win_times"},inplace = True)
1

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,)
1
2
3
4

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




<Figure size 2000x1000 with 0 Axes>
1
2
3
4
5
6

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"])
1

win_lose_times.sort_values(by = "wp",ascending=False)
1

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()
1
2
3
4
5
6
7
8
9

df_player.head()
1

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"]
1
2

df_james.info()
1

<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
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31

#处理一下空值
pd.set_option("display.max_columns",None)
df_james.loc[df_james["FT_PCT"].isnull() == True,:]
1
2
3

df_james.fillna(value = 0,inplace = True)
1

df_james.info()
1

<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
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31

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()
1
2
3
4
5
6
7
8
9
10

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)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23

'胜利场次：23,失败场次：29,胜率：0.44'
1

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)
1
2
3
4
5

27
1

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
1
2
3
4

0.4444444444444444
1

lose_games
1

#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])
1
2
3
4
5
6

#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
1
2
3
4
5

0.6346153846153846
1

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…