【python数据分析】零售场景(附数据集链接)

---

部分词典

列名	对照
Recency	距离最后一次到店的时间
Kidhome	家里有多少个儿童
NumStorePurchases	到店购买数量
AcceptedCmp3	参加活动3
Response	参加最后一次活动

注：对于定性变量，1为是，0为否

前言

数据集github链接
总体目标:你是一名营销分析师，首席营销官告诉你，最近的营销活动没有预期的那么有效。您需要分析数据集来理解这个问题，并提出数据驱动的解决方案。)

---

一、探索性数据分析

导入库，读取文件

import pandas as pd
import seaborn as sns
import numpy as np
from matplotlib import pyplot as plt 
df = pd.read_csv('ml_project1_data.csv')

1.是否存在空值或异常值，如何处理

df.isna().sum()

输出：

ID                      0
Year_Birth              0
Education               0
Marital_Status          0
Income                 24
Kidhome                 0
Teenhome                0
Dt_Customer             0
Recency                 0
MntWines                0
MntFruits               0
MntMeatProducts         0
MntFishProducts         0
MntSweetProducts        0
MntGoldProds            0
NumDealsPurchases       0
NumWebPurchases         0
NumCatalogPurchases     0
NumStorePurchases       0
NumWebVisitsMonth       0
AcceptedCmp3            0
AcceptedCmp4            0
AcceptedCmp5            0
AcceptedCmp1            0
AcceptedCmp2            0
Complain                0
Z_CostContact           0
Z_Revenue               0
Response                0
dtype: int64

Income列有24个空值，先看看它的分布再决定如何处理空值

sns.boxplot(data=df,y='Income')

输出：

有几个异常大的值，为了不改变数据分布所以这里采用中位数填充空值

df['Income'] = df['Income'].fillna(df['Income'].median())

2.有变量需要转换数据类型吗

先看一下数据类型

df.info()

输出：

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 2240 entries, 0 to 2239
Data columns (total 29 columns):
 #   Column               Non-Null Count  Dtype  
---  ------               --------------  -----  
 0   ID                   2240 non-null   int64  
 1   Year_Birth           2240 non-null   int64  
 2   Education            2240 non-null   object 
 3   Marital_Status       2240 non-null   object 
 4   Income               2216 non-null   float64
 5   Kidhome              2240 non-null   int64  
 6   Teenhome             2240 non-null   int64  
 7   Dt_Customer          2240 non-null   object 
 8   Recency              2240 non-null   int64  
 9   MntWines             2240 non-null   int64  
 10  MntFruits            2240 non-null   int64  
 11  MntMeatProducts      2240 non-null   int64  
 12  MntFishProducts      2240 non-null   int64  
 13  MntSweetProducts     2240 non-null   int64  
 14  MntGoldProds         2240 non-null   int64  
 15  NumDealsPurchases    2240 non-null   int64  
 16  NumWebPurchases      2240 non-null   int64  
 17  NumCatalogPurchases  2240 non-null   int64  
 18  NumStorePurchases    2240 non-null   int64  
 19  NumWebVisitsMonth    2240 non-null   int64  
 20  AcceptedCmp3         2240 non-null   int64  
 21  AcceptedCmp4         2240 non-null   int64  
 22  AcceptedCmp5         2240 non-null   int64  
 23  AcceptedCmp1         2240 non-null   int64  
 24  AcceptedCmp2         2240 non-null   int64  
 25  Complain             2240 non-null   int64  
 26  Z_CostContact        2240 non-null   int64  
 27  Z_Revenue            2240 non-null   int64  
 28  Response             2240 non-null   int64  
dtypes: float64(1), int64(25), object(3)
memory usage: 507.6+ KB

有三个object类型的列，下面看看它们分别是什么类型的数据

df[['Education','Marital_Status','Dt_Customer']].head()

输出：

	Education	Marital_Status	Dt_Customer
0	Graduation	Single	2012-09-04
1	Graduation	Single	2014-03-08
2	Graduation	Together	2013-08-21
3	Graduation	Together	2014-02-10
4	PhD	Married	2014-01-19

为方便处理，前两列需要map为数字，时间类就转为datetime类
不过再map之前需要先看下前两列的分布

f, axes = plt.subplots(1, 2, figsize=(16, 4),sharey=True)
for ax,col in zip(axes,['Education','Marital_Status']):
    sns.countplot(x=df[col],ax=ax)

输出：

Marital_Status有三个明显异常值，先看看这些数据正不正常

df[df['Marital_Status'].isin(['Alone','Absurd','YOLO'])]

输出：

	ID	Year_Birth	Education	Marital_Status	Income	Kidhome	Teenhome	Dt_Customer	Recency	MntWines	...	NumWebVisitsMonth	AcceptedCmp3	AcceptedCmp4	AcceptedCmp5	AcceptedCmp1	AcceptedCmp2	Complain	Z_CostContact	Z_Revenue	Response
131	433	1958	Master	Alone	61331.0	1	1	2013-03-10	42	534	...	8	0	0	0	0	0	0	3	11	0
138	7660	1973	PhD	Alone	35860.0	1	1	2014-05-19	37	15	...	5	1	0	0	0	0	0	3	11	1
153	92	1988	Graduation	Alone	34176.0	1	0	2014-05-12	12	5	...	6	0	0	0	0	0	0	3	11	0
2093	7734	1993	Graduation	Absurd	79244.0	0	0	2012-12-19	58	471	...	1	0	0	1	1	0	0	3	11	1
2134	4369	1957	Master	Absurd	65487.0	0	0	2014-01-10	48	240	...	2	0	0	0	0	0	0	3	11	0
2177	492	1973	PhD	YOLO	48432.0	0	1	2012-10-18	3	322	...	8	0	0	0	0	0	0	3	11	0
2202	11133	1973	PhD	YOLO	48432.0	0	1	2012-10-18	3	322	...	8	0	0	0	0	0	0	3	11	1

7 rows × 29 columns

婚姻状态值为Alone或YOLO的数据显示这些顾客家中都有儿女，因而可以将其归为Divorced类，而值为Absurd的顾客家中无儿女，这里简单的将其归为Single类
可以看到婚姻状况为YOLO两行数据除了ID列外其余列的数据相同,因此认定为重复项，因为数据集还可能有其它重复项，所以在这里先简单处理一下

print('除去',len(df)-len(df.drop_duplicates(subset=df.columns[1:])),'重复项')
df=df.drop_duplicates(subset=df.columns[1:])

输出：

除去 182 重复项

现在回头map前两列数据以及把时间列转为datetime类型

df['Marital_Status']=df['Marital_Status'].map({
   'Single':'1','Together':'4','Married':'5',
                                               'Divorced':'2','Widow':'3','Alone':'2',
                                               'Absurd':'1','YOLO':'2'}).astype(int)
df['Education']=df['Education'].map({
   'Graduation':'2', 'PhD':'5', 'Master':'4', 
                                     'Basic':'1', '2n Cycle':'3'}).astype(int)
df['Dt_Customer']=pd.to_datetime(df['Dt_Customer'])

注：数值大小根据来源为生活经验

3.你注意到数据中有什么异常吗?

现在看看其余列的数据分布是否有异常

df.drop(['ID','Education','Marital_Status','Dt_Customer'],1).plot(subplots=True, kind='box',
                                                                  layout=(5,5), figsize=(16,16), patch_artist=True)

输出：