Groupby、value counts和calculate percentage in Pandas

Question

我有groupby state，value计算数据帧的industry。

df.loc[df['state'].isin(['Alabama','Arizona'])].groupby(df['state'])['industry'].value_counts(sort = True)

输出：

state    industry                              
Alabama  Financial Services                        224
         Education                                   7
         Healthcare, Pharmaceuticals, & Biotech      5
         Business Services                           2
         Other                                       2
         Retail                                      2
         Government                                  1
         Manufacturing                               1
         Transportation & Storage                    1
Arizona  Healthcare, Pharmaceuticals, & Biotech     19
         Other                                      13
         Education                                   5
         Retail                                      5
         Transportation & Storage                    5
         Manufacturing                               4
         Travel, Recreation, and Leisure             4
         Consumer Services                           3
         Energy & Utilities                          2
         Financial Services                          2
         Government                                  2
         Business Services                           1
         Computers & Electronics                     1
         Software & Internet                         1
Name: industry, dtype: int64

现在我想更进一步，获得值计数的百分比，例如，对于Alabama，我想知道由224/ (224 + 7 + ... + 1)计算的Financial Services的百分比，等等。

如何通过使用新代码或修改上面的代码来做到这一点？谢谢。

Answer 1

添加normalize

df.loc[df['state'].isin(['Alabama','Arizona'])].groupby(df['state'])['industry'].value_counts(sort = True, normalize=True)