AB测试实战（二）

import numpy as np
import pandas as pd
import scipy.stats as stats
import statsmodels.stats.api as sms
import matplotlib as mpl
import matplotlib.pyplot as plt
import seaborn as sns
from math import ceil

%matplotlib inline


effect_size = sms.proportion_effectsize(0.13, 0.15)   

required_n = sms.NormalIndPower().solve_power(
    effect_size,
    power=0.8, 
    alpha=0.05, 
    ratio=1  
    )                                                  
#对结果向上取整
required_n = ceil(required_n)                         

print(required_n)

df = pd.read_csv('data/ab_data.csv')
df.head()

df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 294478 entries, 0 to 294477
Data columns (total 5 columns):
 #   Column        Non-Null Count   Dtype 
---  ------        --------------   ----- 
 0   user_id       294478 non-null  int64 
 1   timestamp     294478 non-null  object
 2   group         294478 non-null  object
 3   landing_page  294478 non-null  object
 4   converted     294478 non-null  int64 
dtypes: int64(2), object(3)
memory usage: 11.2+ MB

df.pivot_table(index = 'group',columns='landing_page',values = 'user_id',aggfunc='count')

session_counts = df['user_id'].value_counts(ascending=False)
multi_users = session_counts[session_counts>1].count()
multi_users

users = session_counts[session_counts < 2].index
df = df[df['user_id'].isin(users)]

control_sample = df[df['group'] == 'control'].sample(n=required_n, random_state=22)
treatment_sample = df[df['group'] == 'treatment'].sample(n=required_n, random_state=22)
# 这里random_state 为随机数种子, 如果也传入22, 那么后续结果会与讲义中一样
ab_test = pd.concat([control_sample, treatment_sample], axis=0)
ab_test.reset_index(drop=True, inplace=True)
ab_test

ab_test.groupby('group')['landing_page'].value_counts()
# landing_page 落地页类型  old_page老页面 new_page 新页面
# control 控制组  treatment 对照组

conversion_rates = ab_test.groupby('group')['converted'].mean().to_frame()
conversion_rates 

conversion_rates.style.format('{:.3f}')

from statsmodels.stats.proportion import proportions_ztest, proportion_confint
control_results = ab_test[ab_test['group'] == 'control']['converted']     #获取对照组是否转化的数据
treatment_results = ab_test[ab_test['group'] == 'treatment']['converted'] #获取实验组是否转化的数据
n_con = control_results.count() # 获取对照组人数
n_treat = treatment_results.count()  # 获取实验组人数
successes = [control_results.sum(), treatment_results.sum()] # 获取实验组和对照组成功转化的人数
nobs = [n_con, n_treat]

z_stat, pval = proportions_ztest(successes, nobs=nobs)  #计算P值
(lower_con, lower_treat), (upper_con, upper_treat) = proportion_confint(successes, nobs=nobs, alpha=0.05) #计算置信区间

print(f'z statistic: {z_stat:.2f}')
print(f'p-value: {pval:.3f}')
print(f'ci 95% for control group: [{lower_con:.3f}, {upper_con:.3f}]')
print(f'ci 95% for treatment group: [{lower_treat:.3f}, {upper_treat:.3f}]')