CODEWITHSUNDEEP
pythonpandas
calvinkid
calvinkid

Reputation: 43

Creating an interaction matrix in python

Can someone tell me how to generate an interaction matrix from a dataframe as follows?

Input:

    cd1 = [True, False, True, False, False]
    cd2 = [False, True, True, False, False]
    cd3 = [True, False, False, True, True]
    df = pd.DataFrame({"cd1":cd1, "cd2": cd2, "cd3": cd3})

I want an output that looks like this 

   cd1 cd2 cd3
cd1 2   1   1
cd2 1   2   0
cd3 1   0   3

Upvotes: 2

Views: 1595

Answers (2)

BENY
BENY

Reputation: 323226

Just one more option matmul from numpy

pd.DataFrame(np.matmul(df.T.astype(int),df.astype(int)),columns=df.columns,index=df.columns)
Out[283]: 
     cd1  cd2  cd3
cd1    2    1    1
cd2    1    2    0
cd3    1    0    3

Upvotes: 1

cs95
cs95

Reputation: 402493

Convert df to int type and compute the matrix dot product:

df.astype(int).T.dot(df)

     cd1  cd2  cd3
cd1    2    1    1
cd2    1    2    0
cd3    1    0    3

Upvotes: 3

Related Questions