task2

数据集有多少男性和女性？

data['sex'].value_counts()

Male      21790
Female    10771
Name: sex, dtype: int64

数据集女性的平均年龄

data[data['sex'] == 'Female']['age'].mean()

36.85823043357163

数据集中德国公民的比例是多少？

data['native-country'].value_counts(normalize=True)['Germany']

0.004207487485028101

年收入超过 50K 和低于 50K 人群年龄的平均值和标准差是多少？

salary1 = data[data['salary'] == '>50K']['age']
salary2 = data[data['salary'] == '<=50K']['age']
print(salary1.mean(),salary1.std())
print(salary2.mean(),salary2.std())

44.24984058155847 10.51902771985177
36.78373786407767 14.020088490824813

年收入超过 50K 的人群是否都接受过高中以上教育？

data[data['salary'] == ">50K"]['education'].unique()

array(['HS-grad', 'Masters', 'Bachelors', 'Some-college', 'Assoc-voc',
       'Doctorate', 'Prof-school', 'Assoc-acdm', '7th-8th', '12th',
       '10th', '11th', '9th', '5th-6th', '1st-4th'], dtype=object)

使用 groupby 和 describe 统计不同种族和性别人群的年龄分布数据。

for (race,sex),mini_data in data.groupby(['race','sex']):
    print(race,sex)
    print(mini_data['age'].describe())

Amer-Indian-Eskimo Female
count    119.000000
mean      37.117647
std       13.114991
min       17.000000
25%       27.000000
50%       36.000000
75%       46.000000
max       80.000000
Name: age, dtype: float64
Amer-Indian-Eskimo Male
count    192.000000
mean      37.208333
std       12.049563
min       17.000000
25%       28.000000
50%       35.000000
75%       45.000000
max       82.000000
Name: age, dtype: float64
Asian-Pac-Islander Female
count    346.000000
mean      35.089595
std       12.300845
min       17.000000
25%       25.000000
50%       33.000000
75%       43.750000
max       75.000000
Name: age, dtype: float64
Asian-Pac-Islander Male
count    693.000000
mean      39.073593
std       12.883944
min       18.000000
25%       29.000000
50%       37.000000
75%       46.000000
max       90.000000
Name: age, dtype: float64
Black Female
count    1555.000000
mean       37.854019
std        12.637197
min        17.000000
25%        28.000000
50%        37.000000
75%        46.000000
max        90.000000
Name: age, dtype: float64
Black Male
count    1569.000000
mean       37.682600
std        12.882612
min        17.000000
25%        27.000000
50%        36.000000
75%        46.000000
max        90.000000
Name: age, dtype: float64
Other Female
count    109.000000
mean      31.678899
std       11.631599
min       17.000000
25%       23.000000
50%       29.000000
75%       39.000000
max       74.000000
Name: age, dtype: float64
Other Male
count    162.000000
mean      34.654321
std       11.355531
min       17.000000
25%       26.000000
50%       32.000000
75%       42.000000
max       77.000000
Name: age, dtype: float64
White Female
count    8642.000000
mean       36.811618
std        14.329093
min        17.000000
25%        25.000000
50%        35.000000
75%        46.000000
max        90.000000
Name: age, dtype: float64
White Male
count    19174.000000
mean        39.652498
std         13.436029
min         17.000000
25%         29.000000
50%         38.000000
75%         49.000000
max         90.000000
Name: age, dtype: float64

统计男性高收入人群中已婚和未婚（包含离婚和分居）人群各自所占数量。

# 未婚
data[(data['sex'] == 'Male') &
     (data['marital-status'].isin(['Never-married',
                                   'Separated', 'Divorced']))]['salary'].value_counts()

<=50K    7423
>50K      658
Name: salary, dtype: int64

# 已婚
data[(data['sex'] == 'Male') &
     (data['marital-status'].str.startswith('Married'))]['salary'].value_counts()

<=50K    7576
>50K     5965
Name: salary, dtype: int64

计算各国超过和低于 50K 人群各自的平均周工作时长。

for (country, salary), sub_df in data.groupby(['native-country', 'salary']):
    print(country, salary, round(sub_df['hours-per-week'].mean(), 2))

? <=50K 40.16
? >50K 45.55
Cambodia <=50K 41.42
Cambodia >50K 40.0
Canada <=50K 37.91
Canada >50K 45.64
China <=50K 37.38
China >50K 38.9
Columbia <=50K 38.68
Columbia >50K 50.0
Cuba <=50K 37.99
Cuba >50K 42.44
Dominican-Republic <=50K 42.34
Dominican-Republic >50K 47.0
Ecuador <=50K 38.04
Ecuador >50K 48.75
El-Salvador <=50K 36.03
El-Salvador >50K 45.0
England <=50K 40.48
England >50K 44.53
France <=50K 41.06
France >50K 50.75
Germany <=50K 39.14
Germany >50K 44.98
Greece <=50K 41.81
Greece >50K 50.62
Guatemala <=50K 39.36
Guatemala >50K 36.67
Haiti <=50K 36.33
Haiti >50K 42.75
Holand-Netherlands <=50K 40.0
Honduras <=50K 34.33
Honduras >50K 60.0
Hong <=50K 39.14
Hong >50K 45.0
Hungary <=50K 31.3
Hungary >50K 50.0
India <=50K 38.23
India >50K 46.48
Iran <=50K 41.44
Iran >50K 47.5
Ireland <=50K 40.95
Ireland >50K 48.0
Italy <=50K 39.62
Italy >50K 45.4
Jamaica <=50K 38.24
Jamaica >50K 41.1
Japan <=50K 41.0
Japan >50K 47.96
Laos <=50K 40.38
Laos >50K 40.0
Mexico <=50K 40.0
Mexico >50K 46.58
Nicaragua <=50K 36.09
Nicaragua >50K 37.5
Outlying-US(Guam-USVI-etc) <=50K 41.86
Peru <=50K 35.07
Peru >50K 40.0
Philippines <=50K 38.07
Philippines >50K 43.03
Poland <=50K 38.17
Poland >50K 39.0
Portugal <=50K 41.94
Portugal >50K 41.5
Puerto-Rico <=50K 38.47
Puerto-Rico >50K 39.42
Scotland <=50K 39.44
Scotland >50K 46.67
South <=50K 40.16
South >50K 51.44
Taiwan <=50K 33.77
Taiwan >50K 46.8
Thailand <=50K 42.87
Thailand >50K 58.33
Trinadad&Tobago <=50K 37.06
Trinadad&Tobago >50K 40.0
United-States <=50K 38.8
United-States >50K 45.51
Vietnam <=50K 37.19
Vietnam >50K 39.2
Yugoslavia <=50K 41.6
Yugoslavia >50K 49.5

# 交叉表
pd.crosstab(data['native-country'], data['salary'],
            values=data['hours-per-week'], aggfunc=np.mean)