Chapter 71

05 train churn model

NotebookPython 3 (ipykernel)35 cells

In the previous session we trained a model for predicting churn and evaluated it. Now let's deploy it

In [10]python · cell 2

python

import pandas as pd
import numpy as np

from sklearn.model_selection import train_test_split
from sklearn.model_selection import KFold

from sklearn.feature_extraction import DictVectorizer
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import roc_auc_score

In [5]python · cell 3

python

df = pd.read_csv('data-week-3.csv')

df.columns = df.columns.str.lower().str.replace(' ', '_')

categorical_columns = list(df.dtypes[df.dtypes == 'object'].index)

for c in categorical_columns:
    df[c] = df[c].str.lower().str.replace(' ', '_')

df.totalcharges = pd.to_numeric(df.totalcharges, errors='coerce')
df.totalcharges = df.totalcharges.fillna(0)

df.churn = (df.churn == 'yes').astype(int)

In [7]python · cell 4

python

df_full_train, df_test = train_test_split(df, test_size=0.2, random_state=1)

In [18]python · cell 5

python

numerical = ['tenure', 'monthlycharges', 'totalcharges']

categorical = [
    'gender',
    'seniorcitizen',
    'partner',
    'dependents',
    'phoneservice',
    'multiplelines',
    'internetservice',
    'onlinesecurity',
    'onlinebackup',
    'deviceprotection',
    'techsupport',
    'streamingtv',
    'streamingmovies',
    'contract',
    'paperlessbilling',
    'paymentmethod',
]

In [13]python · cell 6

python

def train(df_train, y_train, C=1.0):
    dicts = df_train[categorical + numerical].to_dict(orient='records')

    dv = DictVectorizer(sparse=False)
    X_train = dv.fit_transform(dicts)

    model = LogisticRegression(C=C, max_iter=1000)
    model.fit(X_train, y_train)
    
    return dv, model

In [14]python · cell 7

python

def predict(df, dv, model):
    dicts = df[categorical + numerical].to_dict(orient='records')

    X = dv.transform(dicts)
    y_pred = model.predict_proba(X)[:, 1]

    return y_pred

In [15]python · cell 8

python

C = 1.0
n_splits = 5

In [16]python · cell 9

python

kfold = KFold(n_splits=n_splits, shuffle=True, random_state=1)

scores = []

for train_idx, val_idx in kfold.split(df_full_train):
    df_train = df_full_train.iloc[train_idx]
    df_val = df_full_train.iloc[val_idx]

    y_train = df_train.churn.values
    y_val = df_val.churn.values

    dv, model = train(df_train, y_train, C=C)
    y_pred = predict(df_val, dv, model)

    auc = roc_auc_score(y_val, y_pred)
    scores.append(auc)

print('C=%s %.3f +- %.3f' % (C, np.mean(scores), np.std(scores)))

Output

C=1.0 0.841 +- 0.008

In [17]python · cell 10

python

scores

Output

[0.8423083263338855,
 0.8450681201165409,
 0.8324061810154525,
 0.8319390707936304,
 0.8522598914373568]

In [18]python · cell 11

python

dv, model = train(df_full_train, df_full_train.churn.values, C=1.0)
y_pred = predict(df_test, dv, model)

y_test = df_test.churn.values
auc = roc_auc_score(y_test, y_pred)
auc

Output

0.8572386167896259

Save the model

In [8]python · cell 13

python

import pickle

In [9]python · cell 14

python

output_file = f'model_C={C}.bin'

In [10]python · cell 15

python

output_file

Output

'model_C=1.0.bin'

In [16]python · cell 16

python

f_out = open(output_file, 'wb') 
pickle.dump((dv, model), f_out)
f_out.close()

In [22]python · cell 17

python

!ls -lh *.bin

Output

-rwxrwxrwx 1 alexey alexey 2.5K Sep 30 14:10 'model_C=1.0.bin'

In [21]python · cell 18

python

with open(output_file, 'wb') as f_out: 
    pickle.dump((dv, model), f_out)

Load the model

In [1]python · cell 20

python

import pickle

In [2]python · cell 21

python

input_file = 'model_C=1.0.bin'

In [4]python · cell 22

python

with open(input_file, 'rb') as f_in: 
    dv, model = pickle.load(f_in)

In [8]python · cell 23

python

model

Output

LogisticRegression(max_iter=1000)

In [27]python · cell 24

python

customer = {
    'gender': 'female',
    'seniorcitizen': 0,
    'partner': 'yes',
    'dependents': 'no',
    'phoneservice': 'no',
    'multiplelines': 'no_phone_service',
    'internetservice': 'dsl',
    'onlinesecurity': 'no',
    'onlinebackup': 'yes',
    'deviceprotection': 'no',
    'techsupport': 'no',
    'streamingtv': 'no',
    'streamingmovies': 'no',
    'contract': 'month-to-month',
    'paperlessbilling': 'yes',
    'paymentmethod': 'electronic_check',
    'tenure': 1,
    'monthlycharges': 29.85,
    'totalcharges': 29.85
}

In [28]python · cell 25

python

X = dv.transform([customer])

In [29]python · cell 26

python

y_pred = model.predict_proba(X)[0, 1]

In [30]python · cell 27

python

print('input:', customer)
print('output:', y_pred)

Output

input: {'gender': 'female', 'seniorcitizen': 0, 'partner': 'yes', 'dependents': 'no', 'phoneservice': 'no', 'multiplelines': 'no_phone_service', 'internetservice': 'dsl', 'onlinesecurity': 'no', 'onlinebackup': 'yes', 'deviceprotection': 'no', 'techsupport': 'no', 'streamingtv': 'no', 'streamingmovies': 'no', 'contract': 'month-to-month', 'paperlessbilling': 'yes', 'paymentmethod': 'electronic_check', 'tenure': 1, 'monthlycharges': 29.85, 'totalcharges': 29.85}
output: 0.5912433520805763

Making requests

In [40]python · cell 29

python

import requests

In [41]python · cell 30

python

url = 'http://localhost:9696/predict'

In [42]python · cell 31

python

customer = {
    'gender': 'female',
    'seniorcitizen': 0,
    'partner': 'yes',
    'dependents': 'no',
    'phoneservice': 'no',
    'multiplelines': 'no_phone_service',
    'internetservice': 'dsl',
    'onlinesecurity': 'no',
    'onlinebackup': 'yes',
    'deviceprotection': 'no',
    'techsupport': 'no',
    'streamingtv': 'no',
    'streamingmovies': 'no',
    'contract': 'two_year',
    'paperlessbilling': 'yes',
    'paymentmethod': 'electronic_check',
    'tenure': 1,
    'monthlycharges': 29.85,
    'totalcharges': 29.85
}

In [ ]python · cell 32

python

response = requests.post(url, json=customer).json()

In [39]python · cell 33

python

response

Output

{'churn': True, 'churn_probability': 0.5133820686195286}

In [27]python · cell 34

python

if response['churn']:
    print('sending email to', 'asdx-123d')

Output

sending email to asdx-123d

In [ ]python · cell 35

python

Chapter 71

05 train churn model

NotebookPython 3 (ipykernel)35 cells

In the previous session we trained a model for predicting churn and evaluated it. Now let's deploy it

In [10]python · cell 2

python

import pandas as pd
import numpy as np

from sklearn.model_selection import train_test_split
from sklearn.model_selection import KFold

from sklearn.feature_extraction import DictVectorizer
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import roc_auc_score

In [5]python · cell 3

python

df = pd.read_csv('data-week-3.csv')

df.columns = df.columns.str.lower().str.replace(' ', '_')

categorical_columns = list(df.dtypes[df.dtypes == 'object'].index)

for c in categorical_columns:
    df[c] = df[c].str.lower().str.replace(' ', '_')

df.totalcharges = pd.to_numeric(df.totalcharges, errors='coerce')
df.totalcharges = df.totalcharges.fillna(0)

df.churn = (df.churn == 'yes').astype(int)

In [7]python · cell 4

python

df_full_train, df_test = train_test_split(df, test_size=0.2, random_state=1)

In [18]python · cell 5

python

numerical = ['tenure', 'monthlycharges', 'totalcharges']

categorical = [
    'gender',
    'seniorcitizen',
    'partner',
    'dependents',
    'phoneservice',
    'multiplelines',
    'internetservice',
    'onlinesecurity',
    'onlinebackup',
    'deviceprotection',
    'techsupport',
    'streamingtv',
    'streamingmovies',
    'contract',
    'paperlessbilling',
    'paymentmethod',
]

In [13]python · cell 6

python

def train(df_train, y_train, C=1.0):
    dicts = df_train[categorical + numerical].to_dict(orient='records')

    dv = DictVectorizer(sparse=False)
    X_train = dv.fit_transform(dicts)

    model = LogisticRegression(C=C, max_iter=1000)
    model.fit(X_train, y_train)
    
    return dv, model

In [14]python · cell 7

python

def predict(df, dv, model):
    dicts = df[categorical + numerical].to_dict(orient='records')

    X = dv.transform(dicts)
    y_pred = model.predict_proba(X)[:, 1]

    return y_pred

In [15]python · cell 8

python

C = 1.0
n_splits = 5

In [16]python · cell 9

python

kfold = KFold(n_splits=n_splits, shuffle=True, random_state=1)

scores = []

for train_idx, val_idx in kfold.split(df_full_train):
    df_train = df_full_train.iloc[train_idx]
    df_val = df_full_train.iloc[val_idx]

    y_train = df_train.churn.values
    y_val = df_val.churn.values

    dv, model = train(df_train, y_train, C=C)
    y_pred = predict(df_val, dv, model)

    auc = roc_auc_score(y_val, y_pred)
    scores.append(auc)

print('C=%s %.3f +- %.3f' % (C, np.mean(scores), np.std(scores)))

Output

C=1.0 0.841 +- 0.008

In [17]python · cell 10

python

scores

Output

[0.8423083263338855,
 0.8450681201165409,
 0.8324061810154525,
 0.8319390707936304,
 0.8522598914373568]

In [18]python · cell 11

python

dv, model = train(df_full_train, df_full_train.churn.values, C=1.0)
y_pred = predict(df_test, dv, model)

y_test = df_test.churn.values
auc = roc_auc_score(y_test, y_pred)
auc

Output

0.8572386167896259

Save the model

In [8]python · cell 13

python

import pickle

In [9]python · cell 14

python

output_file = f'model_C={C}.bin'

In [10]python · cell 15

python

output_file

Output

'model_C=1.0.bin'

In [16]python · cell 16

python

f_out = open(output_file, 'wb') 
pickle.dump((dv, model), f_out)
f_out.close()

In [22]python · cell 17

python

!ls -lh *.bin

Output

-rwxrwxrwx 1 alexey alexey 2.5K Sep 30 14:10 'model_C=1.0.bin'

In [21]python · cell 18

python

with open(output_file, 'wb') as f_out: 
    pickle.dump((dv, model), f_out)

Load the model

In [1]python · cell 20

python

import pickle

In [2]python · cell 21

python

input_file = 'model_C=1.0.bin'

In [4]python · cell 22

python

with open(input_file, 'rb') as f_in: 
    dv, model = pickle.load(f_in)

In [8]python · cell 23

python

model

Output

LogisticRegression(max_iter=1000)

In [27]python · cell 24

python

customer = {
    'gender': 'female',
    'seniorcitizen': 0,
    'partner': 'yes',
    'dependents': 'no',
    'phoneservice': 'no',
    'multiplelines': 'no_phone_service',
    'internetservice': 'dsl',
    'onlinesecurity': 'no',
    'onlinebackup': 'yes',
    'deviceprotection': 'no',
    'techsupport': 'no',
    'streamingtv': 'no',
    'streamingmovies': 'no',
    'contract': 'month-to-month',
    'paperlessbilling': 'yes',
    'paymentmethod': 'electronic_check',
    'tenure': 1,
    'monthlycharges': 29.85,
    'totalcharges': 29.85
}

In [28]python · cell 25

python

X = dv.transform([customer])

In [29]python · cell 26

python

y_pred = model.predict_proba(X)[0, 1]

In [30]python · cell 27

python

print('input:', customer)
print('output:', y_pred)

Output

input: {'gender': 'female', 'seniorcitizen': 0, 'partner': 'yes', 'dependents': 'no', 'phoneservice': 'no', 'multiplelines': 'no_phone_service', 'internetservice': 'dsl', 'onlinesecurity': 'no', 'onlinebackup': 'yes', 'deviceprotection': 'no', 'techsupport': 'no', 'streamingtv': 'no', 'streamingmovies': 'no', 'contract': 'month-to-month', 'paperlessbilling': 'yes', 'paymentmethod': 'electronic_check', 'tenure': 1, 'monthlycharges': 29.85, 'totalcharges': 29.85}
output: 0.5912433520805763

Making requests

In [40]python · cell 29

python

import requests

In [41]python · cell 30

python

url = 'http://localhost:9696/predict'

In [42]python · cell 31

python

customer = {
    'gender': 'female',
    'seniorcitizen': 0,
    'partner': 'yes',
    'dependents': 'no',
    'phoneservice': 'no',
    'multiplelines': 'no_phone_service',
    'internetservice': 'dsl',
    'onlinesecurity': 'no',
    'onlinebackup': 'yes',
    'deviceprotection': 'no',
    'techsupport': 'no',
    'streamingtv': 'no',
    'streamingmovies': 'no',
    'contract': 'two_year',
    'paperlessbilling': 'yes',
    'paymentmethod': 'electronic_check',
    'tenure': 1,
    'monthlycharges': 29.85,
    'totalcharges': 29.85
}

In [ ]python · cell 32

python

response = requests.post(url, json=customer).json()

In [39]python · cell 33

python

response

Output

{'churn': True, 'churn_probability': 0.5133820686195286}

In [27]python · cell 34

python

if response['churn']:
    print('sending email to', 'asdx-123d')

Output

sending email to asdx-123d

In [ ]python · cell 35

python