Insurance Customer Segmentation Project Summary

Project Overview

Develop a predictive model for an Insurance company to determine which of their existing Health Insurance policyholders are likely to be interested in a new Vehicle Insurance policy. The model will be built using PyTorch and Scikit.

Business Context

• The client is an Insurance company that currently provides Health Insurance to its customers.
• They want to predict which of their existing policyholders from the past year might be interested in a new Vehicle Insurance policy.
• This prediction will help the company tailor its communication strategy and optimize its business model and revenue.

Data Source

Kaggle Dataset: Health Insurance Cross Sell Prediction

Data Description

The dataset contains information about insurance policyholders, including:

Training Data

Variable	Definition
id	Unique ID for the customer
Gender	Gender of the customer
Age	Age of the customer
Driving_License	0: Customer does not have DL, 1: Customer has DL
Region_Code	Unique code for the customer's region
Previously_Insured	1: Customer has Vehicle Insurance, 0: Customer doesn't
Vehicle_Age	Age of the Vehicle
Vehicle_Damage	1: Vehicle damaged in the past, 0: No damage
Annual_Premium	Amount customer pays as premium per year
Policy_Sales_Channel	Anonymized code for customer outreach channel
Vintage	Number of days customer associated with the company
Response	1: Customer interested, 0: Customer not interested (target variable)

Project Steps

1. Data Preparation: Load and preprocess data using Pandas
2. Model Development: Create a simple machine-learning model using PyTorch
3. Model Training: Train the model on the preprocessed data
4. Inference : Use the trained model and determine binary classification

Key Skills Demonstrated

• PyTorch for neural network development
• Binary classification modeling
• Data preprocessing with Pandas
• Business interpretation of model results

This project showcases the ability to deploy a machine learning solution using PyTorch and Scikit, demonstrating skills in data preprocessing, model development, and cloud deployment in an insurance context.

Development Environment

• Jupyter Notebook for all stages of development
• Local development for initial stages
• AWS SageMaker Notebook Instance for later stages and deployment

Project Goal

The primary goal is to build a model that accurately predicts which existing Health Insurance policyholders are likely to be interested in Vehicle Insurance. This will enable the company to:

1. Target their marketing efforts more effectively
2. Optimize their communication strategy
3. Potentially increase their revenue from Vehicle Insurance policies

Note on Project Evaluation

All code and thought processes will be documented in Jupyter Notebooks to provide a clear, step-by-step demonstration of the project development.