Python数据分析-FIFA2018球员数据分析

#!/usr/bin/env python
# coding: utf-8

# TI=FIFA2018球员数据分析
# 明确分析目的
# 	运动员数量前十名的国家，以及平均身价
# 	各大联赛运动员数量，以及球员平均身价
# 	各俱乐部的平均周薪
# 	英超联赛English Premier League各个俱乐部球员的平均周薪
# 	球员年龄分布情况，不同年龄段平均身价分布

# 引入使用的库
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

# 加载数据文件
df = pd.read_csv('./FIFA_2018_player.csv')

# 查看数据具有哪些列，什么类型
df.info()
# 可见共17994行，league和club有缺失值

RangeIndex: 17994 entries, 0 to 17993
Data columns (total 12 columns):
ID             17994 non-null int64
name           17994 non-null object
full_name      17994 non-null object
nationality    17994 non-null object
league         17741 non-null object
club           17741 non-null object
age            17994 non-null int64
birth_date     17994 non-null object
height_cm      17994 non-null float64
weight_kg      17994 non-null float64
eur_value      17994 non-null float64
eur_wage       17994 non-null float64
dtypes: float64(4), int64(2), object(6)
memory usage: 1.6+ MB

df.head()

	ID	name	full_name	nationality	league	club	age	birth_date	height_cm	weight_kg	eur_value	eur_wage
0	20801	Cristiano Ronaldo	C. Ronaldo dos Santos Aveiro	Portugal	Spanish Primera División	Real Madrid CF	32	1985-02-05	185.0	80.0	95500000.0	565000.0
1	158023	L. Messi	Lionel Messi	Argentina	Spanish Primera División	FC Barcelona	30	1987-06-24	170.0	72.0	105000000.0	565000.0
2	190871	Neymar	Neymar da Silva Santos Jr.	Brazil	French Ligue 1	Paris Saint-Germain	25	1992-02-05	175.0	68.0	123000000.0	280000.0
3	176580	L. Suárez	Luis Suárez	Uruguay	Spanish Primera División	FC Barcelona	30	1987-01-24	182.0	86.0	97000000.0	510000.0
4	167495	M. Neuer	Manuel Neuer	Germany	German Bundesliga	FC Bayern Munich	31	1986-03-27	193.0	92.0	61000000.0	230000.0

df.describe()

	ID	age	height_cm	weight_kg	eur_value	eur_wage
count	17994.000000	17994.000000	17994.000000	17994.000000	1.799400e+04	17994.000000
mean	207791.796543	25.120151	181.271980	75.400856	2.370511e+06	11503.834612
std	32328.527723	4.617428	6.690392	6.994824	5.347250e+06	23050.661073
min	16.000000	16.000000	155.000000	49.000000	0.000000e+00	0.000000
25%	192621.250000	21.000000	177.000000	70.000000	3.000000e+05	2000.000000
50%	214186.000000	25.000000	181.000000	75.000000	7.000000e+05	4000.000000
75%	231615.750000	28.000000	186.000000	80.000000	2.000000e+06	12000.000000
max	241489.000000	47.000000	205.000000	110.000000	1.230000e+08	565000.000000

df.count()
# 可见league 和 club有缺失值

ID             17994
name           17994
full_name      17994
nationality    17994
league         17741
club           17741
age            17994
birth_date     17994
height_cm      17994
weight_kg      17994
eur_value      17994
eur_wage       17994
dtype: int64

# 对于本次的分析目的，其实在加载数据时就可以只加载部分列
# 选出部分列 ID nationality league club age eur_value eur_wage
# 分析的是FIFA2018的数据，age按当年数据计算，birth_date省略
df = df[['ID', 'nationality', 'league', 'club', 'age', 'eur_value', 'eur_wage']]
df

	ID	nationality	league	club	age	eur_value	eur_wage
0	20801	Portugal	Spanish Primera División	Real Madrid CF	32	95500000.0	565000.0
1	158023	Argentina	Spanish Primera División	FC Barcelona	30	105000000.0	565000.0
2	190871	Brazil	French Ligue 1	Paris Saint-Germain	25	123000000.0	280000.0
3	176580	Uruguay	Spanish Primera División	FC Barcelona	30	97000000.0	510000.0
4	167495	Germany	German Bundesliga	FC Bayern Munich	31	61000000.0	230000.0
...	...	...	...	...	...	...	...
17989	237463	England	English League One	Scunthorpe United	17	50000.0	1000.0
17990	11728	England	English League Two	Wycombe Wanderers	47	0.0	1000.0
17991	231381	Scotland	English League Two	Swindon Town	17	60000.0	1000.0
17992	238813	England	English League Two	Crewe Alexandra	18	60000.0	1000.0
17993	238308	Ghana	English League One	Scunthorpe United	18	50000.0	1000.0

17994 rows × 7 columns

df[df.league.isnull()]

	ID	nationality	league	club	age	eur_value	eur_wage
163	188152	Brazil	NaN	NaN	25	0.0	0.0
168	184826	Portugal	NaN	NaN	28	0.0	0.0
271	177413	Belgium	NaN	NaN	28	0.0	0.0
480	176733	Sweden	NaN	NaN	30	0.0	0.0
494	169195	Brazil	NaN	NaN	29	0.0	0.0
...	...	...	...	...	...	...	...
17267	234509	India	NaN	NaN	29	0.0	0.0
17486	234508	India	NaN	NaN	20	0.0	0.0
17489	223760	India	NaN	NaN	24	0.0	0.0
17511	233526	India	NaN	NaN	22	0.0	0.0
17568	231057	New Zealand	NaN	NaN	20	0.0	0.0

253 rows × 7 columns

# 可以看到联赛、俱乐部是空值，同时这些条目的身价、周薪都是0
# 删除数据一般在后面进行，但四行数据都异常可以先删除
df.drop(df[df.league.isnull()].index,inplace=True)

# 查看删除后情况
df.count()

ID             17741
nationality    17741
league         17741
club           17741
age            17741
eur_value      17741
eur_wage       17741
dtype: int64

# 查看数据中数值、浮点型数据整体信息
df.describe()

	ID	age	eur_value	eur_wage
count	17741.000000	17741.000000	1.774100e+04	17741.000000
mean	207756.835522	25.088552	2.404317e+06	11667.887943
std	32421.331072	4.616413	5.377693e+06	23173.181633
min	16.000000	16.000000	0.000000e+00	1000.000000
25%	192621.000000	21.000000	3.250000e+05	2000.000000
50%	214175.000000	25.000000	7.000000e+05	4000.000000
75%	231624.000000	28.000000	2.100000e+06	12000.000000
max	241489.000000	47.000000	1.230000e+08	565000.000000

# 经查看，eur_value 在最小值上有问题，0.000000e+00
# 筛选一下数据，查看eur_value == 0.000000e+00的有多少
df[df['eur_value'] == 0.000000e+00].count()

ID             6
nationality    6
league         6
club           6
age            6
eur_value      6
eur_wage       6
dtype: int64

# eur_value == 0.000000e+00的有6条
df[df['eur_value'] == 0.000000e+00]

	ID	nationality	league	club	age	eur_value	eur_wage
2199	3665	France	French Ligue 1	ES Troyes AC	40	0.0	16000.0
3105	17605	Belgium	Belgian First Division A	Club Brugge KV	40	0.0	14000.0
3272	176900	Colombia	Colombian Primera A	Asociacion Deportivo Cali	40	0.0	2000.0
7734	148745	Norway	Norwegian Eliteserien	Sogndal	41	0.0	2000.0
17628	149727	England	Rep. Ireland Premier Division	St. Patrick's Athletic	37	0.0	1000.0
17990	11728	England	English League Two	Wycombe Wanderers	47	0.0	1000.0

# 使用平均值填充这些身价为0的数据
# df[df['eur_value'] == 0.000000e+00].loc[:,'eur_value'] = 2.404317e+06 警告

# 使用平均值填充这些身价为0的数据
df['eur_value'].replace(0, df['eur_value'].mean(), inplace = True)

df.describe()
# 可见数据已经填充成功

	ID	age	eur_value	eur_wage
count	17741.000000	17741.000000	1.774100e+04	17741.000000
mean	207756.835522	25.088552	2.405130e+06	11667.887943
std	32421.331072	4.616413	5.377511e+06	23173.181633
min	16.000000	16.000000	1.000000e+04	1000.000000
25%	192621.000000	21.000000	3.250000e+05	2000.000000
50%	214175.000000	25.000000	7.000000e+05	4000.000000
75%	231624.000000	28.000000	2.100000e+06	12000.000000
max	241489.000000	47.000000	1.230000e+08	565000.000000

# 检查是否有整个条目完全重复值，若有则展示
df[df.duplicated()]

	ID	nationality	league	club	age	eur_value	eur_wage

# 检查指定列是否有重复值
df[df['ID'].duplicated()]

	ID	nationality	league	club	age	eur_value	eur_wage

# 查看分类统计值是否有不合逻辑的类名
df['league'].value_counts()

Argentinian Superliga            780
English Championship             717
English League One               668
English Premier League           654
Spanish Segunda División         637
English League Two               633
Italian Serie B                  625
USA Major League Soccer          625
Spanish Primera División         602
French Ligue 1                   598
Italian Serie A                  559
Colombian Primera A              552
French Ligue 2                   543
German Bundesliga                537
Japanese J1 League               519
Mexican Liga MX                  518
German 3. Liga                   515
German 2. Bundesliga             510
Portuguese Primeira Liga         509
Turkish Süper Lig                502
Holland Eredivisie               488
Russian Premier League           449
Belgian First Division A         436
Polish Ekstraklasa               418
Saudi Professional League        411
Norwegian Eliteserien            393
Swedish Allsvenskan              389
Danish Superliga                 365
Korean K League Classic          336
Scottish Premiership             321
Chilian Primera División         320
Campeonato Brasileiro Série A    320
Rep. Ireland Premier Division    288
Swiss Super League               263
Austrian Bundesliga              259
Australian A-League              236
Greek Super League               111
South African PSL                 56
Czech Liga                        28
Finnish Veikkausliiga             27
Ukrainian Premier League          24
Name: league, dtype: int64

# 数据清洗完毕，开始分析

# 样本总数
df.count()

ID             17741
nationality    17741
league         17741
club           17741
age            17741
eur_value      17741
eur_wage       17741
dtype: int64

# 数值类型列统计学指标
df.describe()

	ID	age	eur_value	eur_wage
count	17741.000000	17741.000000	1.774100e+04	17741.000000
mean	207756.835522	25.088552	2.405130e+06	11667.887943
std	32421.331072	4.616413	5.377511e+06	23173.181633
min	16.000000	16.000000	1.000000e+04	1000.000000
25%	192621.000000	21.000000	3.250000e+05	2000.000000
50%	214175.000000	25.000000	7.000000e+05	4000.000000
75%	231624.000000	28.000000	2.100000e+06	12000.000000
max	241489.000000	47.000000	1.230000e+08	565000.000000

# 运动员数量前十名的国家
nationality_data = df.groupby('nationality', as_index = False)  #拿出按国家分组的数据
nat_count = nationality_data.count()[['nationality','ID']]     #计数,拿出国家和ID两列
nat_count.rename(columns = {'ID':'ath_count'}, inplace = True) #对列名重命名
nat_head10 = nat_count.sort_values('ath_count', ascending = False).head(10)#降序排序，取前十
nat_head10

	nationality	ath_count
44	England	1631
57	Germany	1147
135	Spain	1020
53	France	966
5	Argentina	962
18	Brazil	803
75	Italy	800
29	Colombia	591
78	Japan	471
105	Netherlands	430

# 运动员数量前十名的国家及其平均球员身价
nat_val_mean = nationality_data[['nationality','eur_value']].mean()
nat_val_mean.rename(columns = {'eur_value':'val_mean'})
nat_head10_val_mean = pd.merge(nat_head10,nat_val_mean, on = 'nationality', how = 'left')
nat_head10_val_mean

	nationality	ath_count	eur_value
0	England	1631	1.425410e+06
1	Germany	1147	2.609010e+06
2	Spain	1020	4.465897e+06
3	France	966	3.314264e+06
4	Argentina	962	2.900120e+06
5	Brazil	803	4.001071e+06
6	Italy	800	2.681325e+06
7	Colombia	591	1.719068e+06
8	Japan	471	8.067091e+05
9	Netherlands	430	3.002930e+06

# 各大联赛运动员数量，以及球员平均身价(操作方式与上述方法类似)
league_data = df.groupby('league', as_index = False)
league_count = league_data.count()[['league','ID']].rename(columns = {'ID':'ath_count'})
league_count.rename(columns = {'ID':'ath_count'}).sort_values('ath_count', ascending = False)
lea_val_mean = league_data[['league','eur_value']].mean().rename(columns = {'eur_value':'val_mean'})
lea_val_mean = pd.merge(league_count, lea_val_mean, on = 'league', how = 'left')
lea_val_mean

	league	ath_count	val_mean
0	Argentinian Superliga	780	1.453788e+06
1	Australian A-League	236	6.848941e+05
2	Austrian Bundesliga	259	7.276062e+05
3	Belgian First Division A	436	1.956719e+06
4	Campeonato Brasileiro Série A	320	2.249016e+06
5	Chilian Primera División	320	2.238234e+06
6	Colombian Primera A	552	9.465567e+05
7	Czech Liga	28	2.141250e+06
8	Danish Superliga	365	7.188767e+05
9	English Championship	717	1.831032e+06
10	English League One	668	4.875075e+05
11	English League Two	633	2.926687e+05
12	English Premier League	654	9.091483e+06
13	Finnish Veikkausliiga	27	2.940741e+05
14	French Ligue 1	598	5.188201e+06
15	French Ligue 2	543	8.015930e+05
16	German 2. Bundesliga	510	1.238333e+06
17	German 3. Liga	515	4.530777e+05
18	German Bundesliga	537	7.702849e+06
19	Greek Super League	111	3.808333e+06
20	Holland Eredivisie	488	2.171250e+06
21	Italian Serie A	559	7.292030e+06
22	Italian Serie B	625	8.344480e+05
23	Japanese J1 League	519	6.472736e+05
24	Korean K League Classic	336	8.854911e+05
25	Mexican Liga MX	518	2.025782e+06
26	Norwegian Eliteserien	393	6.056217e+05
27	Polish Ekstraklasa	418	7.070096e+05
28	Portuguese Primeira Liga	509	3.506257e+06
29	Rep. Ireland Premier Division	288	1.683310e+05
30	Russian Premier League	449	2.679788e+06
31	Saudi Professional League	411	8.512287e+05
32	Scottish Premiership	321	9.114486e+05
33	South African PSL	56	1.150893e+06
34	Spanish Primera División	602	9.257550e+06
35	Spanish Segunda División	637	1.508854e+06
36	Swedish Allsvenskan	389	6.417095e+05
37	Swiss Super League	263	1.134202e+06
38	Turkish Süper Lig	502	2.961036e+06
39	USA Major League Soccer	625	1.484424e+06
40	Ukrainian Premier League	24	8.283750e+06

# 各俱乐部的平均周薪
club_data = df.groupby('club', as_index = False)
club_wage_mean = club_data.mean()[['club','eur_wage']]
club_wage_mean.rename(columns = {'eur_wage':'wage_mean'},inplace = True)
club_wage_mean.sort_values('wage_mean', ascending = False, inplace = True)
club_wage_mean

	club	wage_mean
219	FC Barcelona	194666.666667
466	Real Madrid CF	170821.428571
222	FC Bayern Munich	123384.615385
330	Juventus	122000.000000
377	Manchester United	109030.303030
...	...	...
97	Bray Wanderers	1000.000000
425	PAOK Thessaloniki	1000.000000
88	Bohemian FC	1000.000000
263	Finn Harps	1000.000000
578	Tigres FC	1000.000000

647 rows × 2 columns

# 英超联赛English Premier League
EPL_data = df[df['league'] == 'English Premier League']
EPL_data.describe()

	ID	age	eur_value	eur_wage
count	654.000000	654.000000	6.540000e+02	654.000000
mean	196333.365443	24.711009	9.091483e+06	57840.978593
std	37997.349392	4.769797	1.222195e+07	50627.145927
min	2147.000000	16.000000	6.000000e+04	2000.000000
25%	183551.250000	20.000000	9.125000e+05	17000.000000
50%	201840.000000	25.000000	5.000000e+06	48000.000000
75%	222604.000000	28.000000	1.137500e+07	82000.000000
max	241384.000000	38.000000	9.050000e+07	325000.000000

# 英超联赛English Premier League各个俱乐部球员的平均周薪
EPL_club = EPL_data.groupby('club', as_index = False)
EPL_club_wage_mean = EPL_club.mean()[['club','eur_wage']]
EPL_club_wage_mean.rename(columns = {'eur_wage':'wage_mean'}, inplace = True)
EPL_club_wage_mean.sort_values('wage_mean', ascending = False, inplace = True)
EPL_club_wage_mean

	club	wage_mean
11	Manchester United	109030.303030
4	Chelsea	105181.818182
10	Manchester City	95787.878788
0	Arsenal	91121.212121
9	Liverpool	83250.000000
6	Everton	76484.848485
16	Tottenham Hotspur	69218.750000
19	West Ham United	61818.181818
13	Southampton	51181.818182
15	Swansea City	43878.787879
8	Leicester City	43875.000000
18	West Bromwich Albion	42516.129032
14	Stoke City	41093.750000
17	Watford	40848.484848
12	Newcastle United	40000.000000
1	Bournemouth	38303.030303
5	Crystal Palace	35181.818182
3	Burnley	33666.666667
2	Brighton & Hove Albion	30454.545455
7	Huddersfield Town	23181.818182

# 球员年龄分布情况，不同年龄段平均身价分布
# ⽣成桶，5岁⼀个分桶，根据上述统计数据可知最⼩16岁，最⼤47
bins = np.arange(15, 50, 5)
bins_data = pd.cut(df['age'], bins)
bin_counts = df['age'].groupby(bins_data).count()
print(bin_counts)
# 可以使用matplotlib粗看一下作图效果，以便在正式出图前作调整
bin_counts.plot(kind='pie')

age
(15, 20]    3300
(20, 25]    6749
(25, 30]    5234
(30, 35]    2192
(35, 40]     258
(40, 45]       7
Name: age, dtype: int64