In [1]:

    

import pandas as pd
import numpy as np
from patsy import dmatrix, dmatrices


    

In [2]:

    

df = pd.DataFrame({'x': np.arange(1, 5), 'y': [2*i for i in np.arange(1, 5)]})


    

In [3]:

    

df


    

    
Out[3]:







  

    

      

      
x
      
y
    
  
  

    

      
0
      
1
      
2
    
    

      
1
      
2
      
4
    
    

      
2
      
3
      
6
    
    

      
3
      
4
      
8
    
  

In [4]:

    

df['z'] = df.x + df.y


    

In [5]:

    

df


    

    
Out[5]:







  

    

      

      
x
      
y
      
z
    
  
  

    

      
0
      
1
      
2
      
3
    
    

      
1
      
2
      
4
      
6
    
    

      
2
      
3
      
6
      
9
    
    

      
3
      
4
      
8
      
12
    
  

In [16]:

    

# Basic design matrix
# Pandas compatibility
# Interaction term is included using x*y
dmat = dmatrix('x + y + x*y', df)
print(dmat)


    

    




[[ 1.  1.  2.  2.]
 [ 1.  2.  4.  8.]
 [ 1.  3.  6. 18.]
 [ 1.  4.  8. 32.]]

In [17]:

    

dmat_df = pd.DataFrame(dmat)


    

In [18]:

    

dmat_df.columns = dmat.design_info.term_names


    

In [19]:

    

dmat_df


    

    
Out[19]:







  

    

      

      
Intercept
      
x
      
y
      
x:y
    
  
  

    

      
0
      
1.0
      
1.0
      
2.0
      
2.0
    
    

      
1
      
1.0
      
2.0
      
4.0
      
8.0
    
    

      
2
      
1.0
      
3.0
      
6.0
      
18.0
    
    

      
3
      
1.0
      
4.0
      
8.0
      
32.0
    
  

In [20]:

    

# Design matrices
y, x = dmatrices('z ~ x + x:y', df)


    

In [23]:

    

y


    

    
Out[23]:





DesignMatrix with shape (4, 1)
   z
   3
   6
   9
  12
  Terms:
    'z' (column 0)

In [22]:

    

# Easy to convert to pandas DataFrame
pd.DataFrame(y)


    

    
Out[22]:







  

    

      

      
0
    
  
  

    

      
0
      
3.0
    
    

      
1
      
6.0
    
    

      
2
      
9.0
    
    

      
3
      
12.0
    
  

In [24]:

    

x


    

    
Out[24]:





DesignMatrix with shape (4, 3)
  Intercept  x  x:y
          1  1    2
          1  2    8
          1  3   18
          1  4   32
  Terms:
    'Intercept' (column 0)
    'x' (column 1)
    'x:y' (column 2)

In [28]:

    

# Easy to convert to pandas DataFrame
pd.DataFrame(x, columns=x.design_info.term_names)


    

    
Out[28]:







  

    

      

      
Intercept
      
x
      
x:y
    
  
  

    

      
0
      
1.0
      
1.0
      
2.0
    
    

      
1
      
1.0
      
2.0
      
8.0
    
    

      
2
      
1.0
      
3.0
      
18.0
    
    

      
3
      
1.0
      
4.0
      
32.0
    
  

In [ ]: