차이 검정 (30대 1인 가구주 성별에 따른소득차이 검정)

In [4]:

import pandas as pd
import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt
from matplotlib import rcParams
from matplotlib import font_manager, rc
%matplotlib inline

 

In [5]:

path = '/usr/local/lib/python3.9/dist-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansMono.ttf'
font_name = font_manager.FontProperties(fname=path).get_name()
rc('font', family = font_name)

 

In [6]:

path = '/content/2020_가구마스터_20230320_21294.csv'
df = pd.read_csv(path, encoding='CP949')

 

In [7]:

columns = ['조사연도', '수도권여부', 'MD제공용_가구고유번호', '가구주_성별코드', '가구주_만연령', '가구원수', '가구주_교육정도_학력코드', '가구주_혼인상태코드', '순자산', '부채', '처분가능소득(보완)[경상소득(보완)-비소비지출(보완)]', '가구주_산업대분류코드', '가구주_직업대분류코드', '입주형태코드']
df = df[columns].copy()
df.rename(columns={'조사연도':'year','수도권여부':'metro','MD제공용_가구고유번호':'id','가구주_성별코드':'sex','가구주_만연령':'age','가구원수':'number','가구주_교육정도_학력코드':'education','가구주_혼인상태코드':'marriage','순자산':'asset','부채':'debt','처분가능소득(보완)[경상소득(보완)-비소비지출(보완)]':'income','가구주_산업대분류코드':'industry','가구주_직업대분류코드':'job','입주형태코드':'house'}, inplace=True)

 

In [8]:

# 30대 1인가구
df1 = df.loc[df['number'].isin([1]) & (df['age']>30) & (df['age']<40) ]
df2 = df1[['sex', 'number', 'age', 'income']]

 

In [11]:

import seaborn as sns
df21 = df2.loc[df2['sex'].isin([1])] # 남
df22 = df2.loc[df2['sex'].isin([2])] # 여

sns.kdeplot(df21['income'], shade=True, label='male', clip=(-1000,20000))
sns.kdeplot(df22['income'], shade=True, label='female', clip=(-1000,20000))
plt.xlabel('10,000 won')
plt.legend()
plt.grid(True)
plt.show()

 

 

<ipython-input-11-c150ecee8d93>:5: FutureWarning: 

`shade` is now deprecated in favor of `fill`; setting `fill=True`.
This will become an error in seaborn v0.14.0; please update your code.

  sns.kdeplot(df21['income'], shade=True, label='male', clip=(-1000,20000))
<ipython-input-11-c150ecee8d93>:6: FutureWarning: 

`shade` is now deprecated in favor of `fill`; setting `fill=True`.
This will become an error in seaborn v0.14.0; please update your code.

  sns.kdeplot(df22['income'], shade=True, label='female', clip=(-1000,20000))

 

 

 

유의성 검정¶

z-통계량 구하기

 

In [12]:

df3 = df2[['income']].groupby(df2['sex']).agg(['mean', 'std', 'count'])
df3

 

Out[12]:

	income
	mean	std	count
sex
1	3098.597156	1947.776724	211
2	2743.759398	1865.068730	133

 

In [13]:

mean = df3[('income', 'mean')]
mean_df = mean[1]-mean[2]
np.round(mean_df, 2)

 

Out[13]:

354.84

 

In [15]:

std = df3[('income', 'std')]
count = df3[('income', 'count')]
se1 = std[1]/np.sqrt(count[1])
se2 = std[1]/np.sqrt(count[1])
tot_se = np.sqrt(se1**2 + se2**2)
np.round(tot_se, 2)

 

Out[15]:

189.63

 

In [16]:

z = mean_df / tot_se
round(z, 2)

 

Out[16]:

1.87

 

In [17]:

import scipy as sp
import scipy.stats

rv = sp.stats.norm(loc=0, scale=1) # 평균 0, 표준편차1, 표준정규분포
np.round(1-rv.cdf(z), 2)

 

Out[17]:

0.03

 

 

p-value = 0.03. 따라서 5%의 유의수준에서 귀무가설(30대 남성 1인 가구 소득과 30대 여성 1인 가구와의 소득차이는 없다)는 기각됨.