Identify and interpret churn drivers

Machine Learning for Marketing in Python

Karolis Urbonas

Head of Analytics & Science, Amazon

Plotting decision tree rules

from sklearn import tree
import graphviz

exported = tree.export_graphviz(
            decision_tree=mytree, 
            out_file=None,
            feature_names=cols, 
            precision=1, 
            class_names=['Not churn','Churn'],
            filled = True)

graph = graphviz.Source(exported)
display(graph)

Interpreting decision tree chart

Decision tree visualization

Logistic regression coefficients

Logistic regression returns beta coefficients
Can be interpreted as change in log-odds of churn associated with 1 unit increase in the feature

Logistic Regression Model

Extracting logistic regression coefficients

Coefficients can be extracted using .coef_ method on fitted Logistic Regression instance
```
logreg.coef_
```

array([[ 0.        ,  0.09784772,  0.        , -0.03935476, -0.82068131,
        -0.41231806, -0.14319622, -0.01746504, -0.41830733,  0.        ,
         0.        ,  0.07138468,  0.        ,  0.        ,  0.        ,
         0.        , -0.41424363, -0.59539021,  0.        ,  0.18846525,
         0.        , -0.90766135,  0.90151342,  0.        ]])

Transforming logistic regression coefficients

Log-odds is difficult to interpret
Solution - calculate exponent of the coefficients
This gives us the change in odds associated with 1 unit increase in the feature

coefficients = pd.concat([pd.DataFrame(train_X.columns),
               pd.DataFrame(np.transpose(logit.coef_))], 
               axis = 1)
coefficients.columns = ['Feature', 'Coefficient']

coefficients['Exp_Coefficient'] = np.exp(coefficients['Coefficient'])

coefficients = coefficients[coefficients['Coefficient']!=0]
print(coefficients.sort_values(by=['Coefficient']))

Meaning of transformed coefficients

Logistic Regression Coefficients

Let's practice!

Machine Learning for Marketing in Python