Insurance-Risk-Analytics-And-Predictive-Modeling

Project Description

This project focuses on an end-to-end insurance risk analytics and predictive modeling solution for AlphaCare Insurance Solutions (ACIS) in South Africa. As a marketing analytics engineer, the primary objective is to transform traditional pricing models into a dynamic, risk-based system by leveraging data-driven insights.

The project encompasses three key phases:

1. Exploratory Data Analysis (EDA): A thorough investigation into ACIS's historical car insurance claim data to understand data quality, distributions, temporal trends, and segment-wise profitability. This phase lays the foundation for all subsequent analytical steps.
2. A/B Hypothesis Testing: Statistical validation of key business assumptions about risk drivers and profit margins across different segments (e.g., provinces, zip codes, demographics). This phase provides data-backed evidence for refining segmentation strategies.
3. Predictive Modeling for Risk and Premium Optimization: Development and evaluation of machine learning models to:
   • Predict Claim Severity: Forecast the financial cost of a claim if it occurs.
   • Predict Claim Probability: Determine the likelihood of a policy resulting in a claim.
   • Direct Premium Prediction: Optimize the calculation of appropriate premiums.

By integrating advanced analytics with business objectives, this project aims to identify low-risk customer segments, optimize marketing strategies, reduce premiums where appropriate, attract new clients, and ultimately enhance ACIS's profitability and competitive advantage.

Table of Contents

• Installation
• Usage
• Project Structure
• Contributing
• License

Installation

Prerequisites

• Python 3.8+ (recommended)
• Git

Steps

1. Clone the repository:

   git clone https://github.com/michaWorku/Insurance-Risk-Analytics-and-Predictive-Modeling.git 
   cd Insurance-Risk-Analytics-and-Predictive-Modeling
   
   

   If you created the project in the current directory:

   # Already in the project root
   
   

2. Create and activate a virtual environment:

   python3 -m venv venv
   source venv/bin/activate # On Windows: .\venv\Scripts\activate
   
   

3. Install dependencies:

   pip install -r requirements.txt
   
   

Usage

This project is primarily driven through Jupyter notebooks located in the notebooks/ directory, which execute the modular Python scripts in src/. Follow the notebooks in sequential order for a comprehensive understanding and reproduction of the analysis:

1. Run EDA Notebook:

   • Navigate to the notebooks/ directory.
   • Open EDA.ipynb to perform exploratory data analysis, data quality checks, and initial summarization.
   • Execute all cells in the notebook. This will also preprocess and save the data to data/processed/.

   # From project root:
   # jupyter notebook notebooks/EDA.ipynb
   
   

2. Run Hypothesis Testing Notebook:

   • Ensure EDA.ipynb has been run first to generate the processed data.
   • Open Hypothesis_Testing.ipynb to conduct A/B hypothesis tests on key risk drivers and profit margins.
   • Execute all cells.

   # From project root:
   # jupyter notebook notebooks/Hypothesis_Testing.ipynb
   
   

3. Run Predictive Modeling Notebook:

   • Ensure EDA.ipynb has been run first.
   • Open Predictive_Modeling.ipynb to build, train, evaluate, and interpret machine learning models for claim severity, claim probability, and premium prediction.
   • Execute all cells.

   # From project root:
   # jupyter notebook notebooks/Predictive_Modeling.ipynb
   
   

For deeper insights, explore the src/ directory which contains the modularized Python code for data loading, preprocessing, analysis strategies, and model implementations.

Project Structure

.
├── .vscode/                 # VSCode specific settings
├── .github/                 # GitHub specific configurations (e.g., Workflows)
│   └── workflows/
│       └── unittests.yml    # CI/CD workflow for tests and linting
├── .gitignore               # Specifies intentionally untracked files to ignore
├── requirements.txt         # Python dependencies
├── pyproject.toml           # Modern Python packaging configuration (PEP 517/621)
├── README.md                # Project overview, installation, usage
├── Makefile                 # Common development tasks (setup, test, lint, clean)
├── .env                     # Environment variables (e.g., API keys - kept out of Git)
├── src/                     # Core source code for the project
│   ├── __init__.py
│   ├── core/                # Core logic/components
│   ├── models/              # Data models, ORM definitions, ML models
│   ├── utils/              # Utility functions, helper classes
│   └── services/            # Business logic, service layer
├── tests/                   # Test suite (unit, integration)
│   ├── unit/                # Unit tests for individual components
│   └── integration/         # Integration tests for combined components
├── notebooks/               # Jupyter notebooks for experimentation, EDA, prototyping
├── scripts/                 # Standalone utility scripts (e.g., data processing, deployment)
├── docs/                    # Project documentation (e.g., Sphinx docs)
├── data/                    # Data storage (raw, processed)
│   ├── raw/                 # Original, immutable raw data
│   └── processed/           # Cleaned, transformed, or feature-engineered data
├── config/                  # Configuration files
└── examples/                # Example usage of the project components

Contributing

Guidelines for contributing to the project.

License

This project is licensed under the MIT License. (Create a LICENSE file if you want to use MIT)