LidetuTadesse/insurance-risk-insights

Risk and predictive analytics for car insurance planning at AlphaCare Insurance Solutions. Includes A/B testing, statistical modeling, and machine learning to identify low risk customers and optimize marketing strategy in South Africa. # Insurance Risk Insights A data science project analyzing auto insurance data to uncover patterns in risk, profitability, and claims using statistical analysis, EDA, and modeling. Built with clean code, Git versioning, and CI/CD practices. --- ## Project Scope - Loss ratio analysis across regions, demographics, and vehicle types - Temporal trends in claims and premiums - Feature engineering and statistical testing - Predictive modeling and risk profiling - Automation via GitHub Actions --- ## Setup ```bash git clone github.com /insurance-risk-insights.git cd insurance-risk-insights python -m venv .venv source .venv/bin/activate # or .venv\Scripts\activate (Windows) pip install -r requirements.txt 🗂️ Structure bash Copy Edit ├── notebooks/ # EDA and prototyping ├── src/ # Core logic and processing ├── scripts/ # Utilities and helpers ├── tests/ # Unit tests ├── .github/ # CI/CD workflows ├── .vscode/ # Dev environment config Tech Stack Python • Pandas • Seaborn • Scikit-learn • Git • GitHub Actions ## 📊 Business Objective The goal is to analyze historical car insurance claims to: - Optimize marketing strategy - Identify low-risk segments for reduced premiums - Recommend data-driven improvements to insurance offerings ## Key Areas of Analysis ### A/B Hypothesis Testing - Risk differences across provinces, zip codes, and gender - Profit margin differences across geographic regions ### Statistical Modeling & Machine Learning - Linear regression per zipcode to predict total claims - ML models to predict optimal premium values based on: - Car features - Owner demographics - Geographic information - Other relevant features ### Insurance Domain Knowledge - Incorporated research on key insurance terminologies ## Methodologies - Data cleaning and preprocessing - Exploratory data analysis (EDA) - Hypothesis testing (t-tests, ANOVA) - Linear regression and predictive modeling - Feature importance an …