计算年份子集中的类别并除以子集中的总数

Question

I am counting the number of negative numbers and positive numbers within each year. Ultimately I want to get the percent of negative and positive for each year.

I tried groupby year and counting the categories, but the new columns appears with no name.

    df1= df.groupby(['Year','Count of Negative/Positive Margins'])['Count of Negative/Positive Margins'].count()

    df1.head()
    Out[194]: 
    Year  Count of Negative/Positive Margins
    2005  1                                     4001
          2                                     1373
    2006  1                                     4046
          2                                     1304
    2007  1                                     4156
    Name: Count of Negative/Positive Margins, dtype: int64 

This my expected output:

    2005  1                                     74%
          2                                     26%
    .
    .
    .

Answer 1

Use SeriesGroupBy.value_counts with grouping only column Year and parameter normalize=True , then multiple by 100 , round by Series.round , convert to strings and add % :

df = (df.groupby('Year')['Count of Negative/Positive Margins']
        .value_counts(normalize=True)
        .mul(100)
        .round()
        .astype(str)
        .add('%')
        .reset_index(name='percentage')
       )