单元格中具有多个值的表的计数/透视

Question

I've got some data that looks like this:

Class                    Instructor
Intro to Philosophy      Jake
Algorithms               Ashley/Jake
Spanish I                Ashley
Vector Calculus          Jake
Intro to Philosophy      Jake

How can I get to a count or pivot that looks like below where the instance that Ashley and Jake both teach a class is properly added to the counts? The instance of one instructor is trivial, but two or more for a single class in the same cell is tripping me up.

I'd like to get to something like this:

                         Jake        Ashley
Intro to Philosophy         2             0
Algorithms                  1             1
Spanish I                   0             1
Vector Calculus             1             0
Total                       4             2

Answer 1

You can use .str.get_dummies to split and binarize the Instructor field. Then you can groupby Class :

ret = (df['Instructor'].str.get_dummies('/')
     .groupby(df['Class']).sum()
)
ret.loc['Total'] = ret.sum()

Output:

                     Ashley  Jake
Class                            
Algorithms                1     1
Intro to Philosophy       0     2
Spanish I                 1     0
Vector Calculus           0     1
Total                     2     4

Answer 2

You can do this:

In [1746]: df.Instructor = df.Instructor.str.split('/')

In [1747]: df = df.explode('Instructor')

In [1751]: x = df.groupby('Instructor').Class.value_counts().reset_index(level=0).pivot(columns='Instructor', values='Class').fillna(0)

In [1754]: x.loc['Total'] = x.sum()

In [1755]: x
Out[1755]: 
Instructor           Ashley  Jake
Class                            
Algorithms              1.0   1.0
Intro_to_Philosophy     0.0   2.0
Spanish_I               1.0   0.0
Vector_Calculus         0.0   1.0
Total                   2.0   4.0

Answer 3

Let us do crosstab after explode

df.Instructor = df.Instructor.str.split('/')

df = df.explode('Instructor')

out = pd.crosstab(df['Class'], df['Instructor'])