Pandas pivot_table using columns names

Question

I have a pandas dataframe that look like this :

ID, tag, score1
A1,  T1,     10
A1,  T1,      0
A1,  T2,     20 
A1,  T2,      0 
A2,  T1,     10
A2,  T1,     10
A2,  T2,     20
A2,  T2,     20

Using pandas pivot_table function, I am able to pivot the table in order to obtain the following dataframe :

df.pivot_table(index= 'tag' , columns='ID', values= 'score1' , aggfunc='mean')

     A1,  A2
T1    5,  10   
T2   10,  20

Now let's say that my input dataframe have multiple score columns :

ID, tag, score1, score2, score3
A1,  T1,     10,    100,   1000
A1,  T1,      0,      0,      0
A1,  T2,     20,    200,   2000    
A1,  T2,      0,      0,      0     
A2,  T1,     10,    100,   1000
A2,  T1,     10,    100,   1000
A2,  T2,     20,    200,   2000
A2,  T2,     20,    200,   2000

And I am looking for a way to pivot the data to get the following result :

df.pivot_table(index= ??? , columns='ID', values= ??? , aggfunc='mean').round(-3)

         A1,   A2
score1  7.5,   15   
score2   75,  150
score3  750, 1500

This time I don't want to pivot using values of a column, but directly using multiple column names.

Is there a way to do this using pivot_table() or am I going in the wrong direction ?

Answer 1

Aggregate mean and then transpose by T :

df = df.groupby('tag').mean().T
print (df)
tag        T1      T2
score1    7.5    15.0
score2   75.0   150.0
score3  750.0  1500.0

Answer 2

Yes, you can use pivot_table like this:

df1.pivot_table(columns='ID', aggfunc='mean')

Output:

ID          A1      A2
 score1    7.5    15.0
 score2   75.0   150.0
 score3  750.0  1500.0