未决问题的记录

本文是个人学习中遇到问题的记录。

---

2018/9/20

q1：多层结构的dataframe怎么理成单层结构的，对其数据进行处理？多层结构的dataframe怎么理成单层结构的，对其数据进行处理？

from pandas_datareader import wb
mathces = wb.search("gdp.*capita.*const")
dat = wb.download(indicator="NY.GDP.PCAP.KD",country=["US","CA","MX","CHN"],start=2015,end=2017)
print(dat)

                    NY.GDP.PCAP.KD
country       year                
Canada        2017    51315.888975
              2016    50407.341330
              2015    50303.836848
China         2017     7329.089299
              2016     6894.464522
              2015     6496.624013
Mexico        2017     9946.157994
              2016     9871.670109
              2015     9717.898430
United States 2017    53128.539700
              2016    52319.163351
              2015    51933.404806

a1: 了解层次化索引的相关处理后，解决本问题

2018/9/22
首先层叠处理一下，长格式转成一列列数据的宽格式

dat2 = dat.unstack(0)
print(dat2)

 NY.GDP.PCAP.KD                                        
country         Canada        China       Mexico United States
year                                                          
2015      50303.836848  6496.624013  9717.898430  51933.404806
2016      50407.341330  6894.464522  9871.670109  52319.163351
2017      51315.888975  7329.089299  9946.157994  53128.539700

然后对数据进行运算

# 计算年变化百分比
returns = dat2.pct_change()
print(returns)

        NY.GDP.PCAP.KD                                  
country         Canada     China    Mexico United States
year                                                    
2015               NaN       NaN       NaN           NaN
2016          0.002058  0.061238  0.015824      0.007428
2017          0.018024  0.063040  0.007546      0.015470

# 计算相关系数
print(returns.corr())

                             NY.GDP.PCAP.KD                           
country                              Canada China Mexico United States
               country                                                
NY.GDP.PCAP.KD Canada                   1.0   1.0   -1.0           1.0
               China                    1.0   1.0   -1.0           1.0
               Mexico                  -1.0  -1.0    1.0          -1.0
               United States            1.0   1.0   -1.0           1.0

# 计算协方差
print(returns.cov())

                             NY.GDP.PCAP.KD                                  
country                              Canada     China    Mexico United States
               country                                                       
NY.GDP.PCAP.KD Canada              0.000127  0.000014 -0.000066      0.000064
               China               0.000014  0.000002 -0.000007      0.000007
               Mexico             -0.000066 -0.000007  0.000034     -0.000033
               United States       0.000064  0.000007 -0.000033      0.000032

---

2018/9/22

q2：numpy 中array生成中，怎么预设缺失值？

arr10 =np.array([[[1,2,3],[4,5]],[[6,7,8],[9,10,11]]],)
print(arr10)

[[list([1, 2, 3]) list([4, 5])]
 [list([6, 7, 8]) list([9, 10, 11])]]
 #本意想生成三维数组
 [[['1' '2' '3']
  ['4' '5' nan]]

 [['6' '7' '8']
  ['9' '10' '11']]]

---

2018/9/25

q2：python for data analysis 第一版第249页中代码运行不出书中结果

party_counts =pd.crosstab(tips.day,tips.size)
party_counts

col_0  1708
day        
Fri      19
Sat      87
Sun      76
Thur     62

a3: 可能是内建函数优先级问题，修改代码后解决

2018/9/27

import pandas as pd
tips = pd.read_csv('tips.csv')
# print(tips[:5])
# print(tips.size)
# party_counts =pd.crosstab(tips.day,tips.size)
party_counts =pd.crosstab(tips["day"],tips["size"])
party_counts2 =pd.crosstab(tips.day,tips["size"])
print(party_counts)
print('===========')
print(party_counts2)

运行结果和书中一致，且注意tips.day 等价于 tips[‘day’]

size  1   2   3   4  5  6
day                      
Fri   1  16   1   1  0  0
Sat   2  53  18  13  1  0
Sun   0  39  15  18  3  1
Thur  1  48   4   5  1  3
===========
size  1   2   3   4  5  6
day                      
Fri   1  16   1   1  0  0
Sat   2  53  18  13  1  0
Sun   0  39  15  18  3  1
Thur  1  48   4   5  1  3