python验证多组数据之间有无显著差异

from statsmodels.fORMula.api import ols
from statsmodels.stats.anova import anova_lm
model = ols('orders~C(label)',data=need_data).fit()
anova_table = anova_lm(model, typ = 2)
print(anova_table)

data2=data1.melt(id_vars=['性别'],value_name='观察频数')
data2.head()

rate=(data2.groupby(['性别'])['观察频数'].sum()/data2.groupby(['性别'])['观察频数'].sum().sum()).reset_index()
rate.columns=['性别','rate']
rate

group_sum=data2.groupby(['组别'])['观察频数'].sum().reset_index()
group_sum.columns=['组别','组内用户数']
group_sum

import math
data3=pd.merge(data2,group_sum,on=['组别'],how='left')
data3=pd.merge(data3,rate,on=['性别'],how='left')
data3['期望频数']=data3['组内用户数']*data3['rate']
data3['卡方值']=data3.apply(lambda x: math.pow((x.期望频数-x.观察频数),2)/x.期望频数,axis=1)
data3.head()