CODEWITHSUNDEEP
pythonpandas
DGraham
DGraham

Reputation: 715

percentage of column pandas

I have a very basic table:

    gov dis value1  value2
0   a   a_1 8   8
1   a   a_2 7   18
2   a   a_3 3   2
3   a   a_4 12  12
4   b   b_1 4   11
5   b   b_2 16  9.....

I can create a basic pivot table from this using:

 t = pd.pivot_table(per_t,index=["gov"],values=['value1'],aggfunc=[np.mean])

To create: 

mean
value1
gov 
a   7.5
b   12.0

I want to convert this into percentages of the column, and have been unable to find anything clear on how to do this? Is there a way of altering the aggfunc to calculate the column percentage?

Upvotes: 2

Views: 1512

Answers (1)

Andy Hayden
Andy Hayden

Reputation: 375565

You can divide by the sum:

In [11]: t.sum()
Out[11]:
mean  value1    17.5
dtype: float64

In [12]: t / t.sum()
Out[12]:
         mean
       value1
gov
a    0.428571
b    0.571429

In [13]: (t / t.sum()) * 100
Out[13]:
          mean
        value1
gov
a    42.857143
b    57.142857

Note: I think you have to do this after the pivot as an aggregation function doesn't have access to the other groups.

Upvotes: 3

Related Questions