dsal1951

| Group | Base Convert Rate | Number of Converters | Number of Leads | Convert Rate | Convert Index |
|-------|-------------------|----------------------|-----------------|--------------|---------------|
| 0     | 0.053             | 82                   | 290             | 0.283        | 5.34x         |
| 1     | 0.053             | 21                   | 292             | 0.072        | 1.36x         |
| 2     | 0.053             | 18                   | 297             | 0.061        | 1.14x         |
| 3     | 0.053             | 6                    | 283             | 0.021        | 0.40x         |
| 4     | 0.053             | 9                    | 338             | 0.027        | 0.50x         |
| 5     | 0.053             | 2                    | 300             | 0.007        | 0.13x         |
| 6     | 0.053             | 3                    | 300             | 0.01         | 0.19x         |
| 7     | 0.053             | 6                    | 300             | 0.02         | 0.38x         |

dsal1951

#Observations where y=1
total_converts = df['CONVERT_ACTUAL'].sum()
#Total Observations
total_leads = df.index.size
baseline_convert_prob = total_converts/float(total_leads)

#Calculate Index of Positive Outcome Rate in Each Decile to the Baseline Positive Rate
lift_df['LIFT_INDEX'] = (lift_df['lift']/baseline_convert_prob)*100

#Add Baseling Positive Rate to DataFrame

dsal1951

group_df = df.groupby('GROUP')

#Total Number of Leads in each Group
leads_count = group_df['CONVERT_ACTUAL'].count()
#Number of Leads who Converted in Each Group
convert_count = group_df['CONVERT_ACTUAL'].sum()

lift = convert_count/leads_count

dsal1951

#Actual Outcome (y=1 converted to customer, y=0 did not convert to customer
convert_actual = y
#Predicted Probability that lead converts into a customer
convert_prob = clf.predict_proba(x)
cols = ['CONVERT_ACTUAL','CONVERT_PROB']
data = [convert_actual,convert_prob[:,1]]
df = pd.DataFrame(dict(zip(cols,data)))

#Sort Ascending based on Predicted Probability y=1
df.sort_values('CONVERT_PROB',ascending=False)

dsal1951

#Create dataset
x,y = datasets.make_classification(n_samples=10000, n_features=20,
                                   n_informative=15,n_redundant=5,
                                   n_classes=2, weights=[0.95,0.05],
                                   random_state=1000)
#Split into Training and Test sets
x_train, x_test, y_train, y_test =  cross_validation.train_test_split(x,y,
                                                                      test_size=0.3,random_state=1000)
#Train a Decision Tree to Predict y                                                                      
clf = tree.DecisionTreeClassifier(min_samples_leaf=10)

dsal1951

def calc_lift(x,y,clf,bins=10):
	"""
	Takes input arrays and trained SkLearn Classifier and returns a Pandas
	DataFrame with the average lift generated by the model in each bin
	
	Parameters
	-------------------
	x:    Numpy array or Pandas Dataframe with shape = [n_samples, n_features]

	y:    A 1-d Numpy array or Pandas Series with shape = [n_samples]

dsal1951

// In app.js or main.js or whatever:
// var myApp = angular.module('askchisne', ['ngSanitize', 'ngAnimate', 'ui.bootstrap', 'ui.bootstrap.tpls']);

// This filter makes the assumption that the input will be in decimal form (i.e. 17% is 0.17).
myApp.filter('percentage', ['$filter', function ($filter) {
  return function (input, decimals) {
    return $filter('number')(input * 100, decimals) + '%';
  };
}]);