from flask import Flask, request, jsonify, render_template, session, url_for, redirect
from flask_wtf import FlaskForm
from flask_wtf import Form
from flask_bootstrap import Bootstrap
from wtforms import StringField, SubmitField, SelectField
import pickle
import joblib
import numpy as np
app = Flask(__name__, template_folder='templates')
app.config['SECRET_KEY'] = b'_5#y2L"F4Q8z\n\xec]/'
model = joblib.load('ensemble_model.pkl')
def return_prediction(model, input_json) :
input_data = [[input_json[k] for k in input_json.keys()]]
prediction = model.predict(input_data)[0]
return prediction
class PredictForm(FlaskForm):
age = StringField("Age")
gender = SelectField("Gender", choices=[('male', 'Male'), ('female', 'Female')])
annual_income = StringField("Annual Income")
marital_status = SelectField("Marital Status", choices=[('married', 'Married'), ('single', 'Single'), ('divorced', 'Divorced')])
dependents = StringField("Number of Dependents")
education_level = SelectField("Education Level", choices=[('high_school', "High School"), ('bachelors', "Bachelor's"), ("masters", "'Master's"), ('phd', "PHD")])
occupation = SelectField("Occupation", choices=[('unemployed', 'Unemployed'), ('self_employed', 'Self-Employed'), ('employed', 'Employed')])
health_score = StringField("Health Score")
location = SelectField("Location", choices=[('rural', 'Rural'), ('suburban', 'Suburban'), ('urban', 'Urban')])
policy_type = SelectField("Policy Type", choices=[('basic', 'Basic'), ('comprehensive', 'Comprehensive'), ('premium', 'Premium')])
previous_claims = StringField("Previous Claims")
vehicle_age = StringField("Vehicle Age (in years)")
credit_score = StringField("Credit Score")
insurance_duration = StringField("Insurance Duration")
customer_feedback = SelectField("Customer Feedback", choices=[('poor', 'Poor'), ('average', 'Average'), ('good', 'Good')])
smoking_status = SelectField("Smoking Status", choices=[('yes', 'Yes'), ('no', 'No')])
exercise_frequency = SelectField("Exercise Frequency", choices=[('rarely', 'Rarely'), ('monthly', 'Monthly'), ('weekly', 'Weekly'), ('daily', 'Daily')])
property_type = SelectField("Property Type", choices=[('apartment', 'Apartment'), ('condo', 'Condo'), ('house', 'House')])
year = StringField("Year")
day = StringField("Day (of the month)")
month = StringField("Month (1-12)")
month_name = ""
day_of_week = StringField("Day of the Week (1-7)")
contract_length = ""
income_per_dependent = ""
credit_score_insurance_duration = ""
health_risk_score = ""
credit_health_score = ""
health_age_interaction = ""
submit = SubmitField("Predict")
@app.route('/', methods=['GET', 'POST'])
def index():
form = PredictForm()
if form.validate_on_submit():
# age
session['age'] = form.age.data
# gender
if(form.gender.data == 'female'):
session['pronoun'] = 0
session['gender'] = 'Female'
else:
session['pronoun'] = 1
session['gender'] = 'Male'
# annual income
session['annual_income'] = form.annual_income.data
# marital status
if(form.marital_status.data == 'married'):
session['married'] = 0
session['marital_status'] = 'Married'
elif(form.marital_status.data == 'single'):
session['married'] = 2
session['marital_status'] = 'Single'
else:
session['married'] = 1
session['marital_status'] = 'Divorced'
# dependents
session['dependents'] = form.dependents.data
# education level
if(form.education_level.data == 'high_school'):
session['education'] = 0
session['education_level'] = 'High School'
elif(form.education_level.data == 'bachelors'):
session['education'] = 1
session['education_level'] = 'Bachelor\'s'
elif(form.education_level.data == 'masters'):
session['education'] = 2
session['education_level'] = 'Master\'s'
else:
session['education'] = 3
session['education_level'] = 'PHD'
# occupation
if(form.occupation.data == 'unemployed'):
session['occupation'] = 0
session['occupation_status'] = 'Unemployed'
elif(form.occupation.data == 'self_employed'):
session['occupation'] = 1
session['occupation_status'] = 'Self-Employed'
else:
session['occupation'] = 2
session['occupation_status'] = 'Employed'
# health score
session['health_score'] = form.health_score.data
# location
if(form.location.data == 'rural'):
session['place'] = 0
session['location'] = 'Rural'
elif(form.location.data == 'suburban'):
session['place'] = 1
session['location'] = 'Suburban'
else:
session['place'] = 2
session['location'] = 'Urban'
# policy type
if(form.policy_type.data == 'basic'):
session['policy_type'] = 0
session['policy'] = 'Basic'
elif(form.policy_type.data == 'comprehensive'):
session['policy_type'] = 1
session['policy'] = 'Comprehensive'
else:
session['policy_type'] = 2
session['policy'] = 'Premium'
# previous claims
session['previous_claims'] = form.previous_claims.data
# vehicle age
session['vehicle_age'] = form.vehicle_age.data
# credit score
session['credit_score'] = form.credit_score.data
# insurance duration
session['insurance_duration'] = form.insurance_duration.data
# customer feedback
if(form.customer_feedback.data == 'poor'):
session['feedback'] = 0
session['feedback_status'] = 'Poor'
elif(form.customer_feedback.data == 'average'):
session['feedback'] = 1
session['feedback_status'] = 'Average'
else:
session['feedback'] = 2
session['feedback_status'] = 'Good'
# smoking status
if(form.smoking_status.data == 'yes'):
session['smoke'] = 1
session['smoking_status'] = 'Yes'
else:
session['smoke'] = 0
session['smoking_status'] = 'No'
# exercise frequency
if(form.exercise_frequency.data == 'rarely'):
session['exercise'] = 0
session['exercise_frequency'] = 'Rarely'
elif(form.exercise_frequency.data == 'monthly'):
session['exercise'] = 1
session['exercise_frequency'] = 'Monthly'
elif(form.exercise_frequency.data == 'weekly'):
session['exercise'] = 2
session['exercise_frequency'] = 'Weekly'
else:
session['exercise'] = 3
session['exercise_frequency'] = 'Daily'
# property type
if(form.property_type.data == 'apartment'):
session['property'] = 0
session['property_type'] = 'Apartment'
elif(form.property_type.data == 'condo'):
session['property'] = 1
session['property_type'] = 'Condo'
else:
session['property'] = 2
session['property_type'] = 'House'
# year
session['year'] = form.year.data
# day
session['day'] = form.day.data
# month
session['month'] = form.month.data
session['month_name'] = form.month.data
# day of the week
session['day_of_week'] = form.day_of_week.data
# contract length
session['contract_length'] = 0
if(float(session['insurance_duration']) < 1):
session['contract_length'] = 0
elif(float(session['insurance_duration']) < 5):
session['contract_length'] = 1
else:
session['contract_length'] = 2
if(float(session['dependents']) == 0):
session['dependents'] = 1
# income per dependent
session['income_per_dependent'] = float(session['annual_income']) / float(session['dependents'])
# credit score insurance duration
session['credit_score_insurance_duration'] = float(session['credit_score']) * float(session['insurance_duration'])
# health risk score
session['health_risk_score'] = float(session['smoke']) + float(session['exercise']) - float(session['health_score']) / 20
# credit health score
session['credit_health_score'] = float(session['credit_score']) * float(session['health_score'])
# health age interaction
session['health_age_interaction'] = float(session['health_score']) * float(session['age'])
return redirect(url_for('predict'))
return render_template('index.html', form=form)
# return """
# Model Deployment
#
Use a POST request to /predict to get a prediction of Premium Amount
#
# - Annual Income
# - Age
# - Gender
# - Dependents
#
# """
@app.route('/predict')
def predict():
content = {}
content['Age'] = float(session['age'])
content['Gender'] = float(session['pronoun'])
content['Annual Income'] = float(session['annual_income'])
content['Marital Status'] = float(session['married'])
content['Number of Dependents'] = float(session['dependents'])
content['Education Level'] = float(session['education'])
content['Occupation'] = float(session['occupation'])
content['Health Score'] = float(session['health_score'])
content['Location'] = float(session['place'])
content['Policy Type'] = float(session['policy_type'])
content['Previous Claims'] = float(session['previous_claims'])
content['Vehicle Age'] = float(session['vehicle_age'])
content['Credit Score'] = float(session['credit_score'])
content['Insurance Duration'] = float(session['insurance_duration'])
content['Customer Feedback'] = float(session['feedback'])
content['Smoking Status'] = float(session['smoke'])
content['Exercise Frequency'] = float(session['exercise'])
content['Property Type'] = float(session['property'])
content['Year'] = float(session['year'])
content['Day'] = float(session['day'])
content['Month'] = float(session['month'])
content['Month Name'] = float(session['month_name'])
content['Day of the Week'] = float(session['day_of_week'])
content['contract length'] = float(session['contract_length'])
content['Income to Dependents Ratio'] = float(session['income_per_dependent'])
content['CreditScore_InsuranceDuration'] = float(session['credit_score_insurance_duration'])
content['Health_Risk_Score'] = float(session['health_risk_score'])
content['Credit_Health_Score'] = float(session['credit_health_score'])
content['Health_Age_Interaction'] = float(session['health_age_interaction'])
prediction = return_prediction(model, content)
prediction = np.expm1(prediction)
return render_template('prediction.html', prediction=prediction)
Bootstrap = Bootstrap(app)
if __name__ == '__main__':
app.run(debug=True)