Motor Insurance Conversion Modelling

This project aims to predict the likelihood of a customer converting a motor insurance quote into an actual policy, and uncover key insights to support marketing and pricing strategy optimization.

---

Project Structure

.
├── data/                   # Input datasets (CSV and Parquet)
│   ├── conversion_data.csv
│   ├── conversion_data.parquet
│   ├── vehicle_classifier.csv
│   └── vehicle_classifier.parquet
├── functions/             # Local modules
│   ├── machine_learning_models.py
│   ├── plot_utils.py
│   └── utils.py
├── models/
│   ├── final_best_model.pkl
│   └── results_df.csv     # Models comparison results
├── Motor Insurance Conversion Modelling.ipynb
└── requirements.txt       # Project dependencies

---

⚙Installation

1. Go to project folder:

   cd motor-insurance-conversion

2. Create and activate a virtual environment (optional but recommended):

   python -m venv .venv
   .venv\Scripts\activate    # On Windows

3. Install required packages:

   pip install -r requirements.txt

---

Running the Project

1. Launch Jupyter Notebook:

   jupyter notebook

2. Open and run: Motor Insurance Conversion Modelling.ipynb

   This notebook includes:

   • Data loading and merging
   • Exploratory Data Analysis
   • Feature engineering and correlation analysis
   • Model training with hyperparameter tuning
   • Evaluation and visualization of model performance
   • Business insights and recommendations

---

Deliverables

• Jupyter Notebook with full pipeline
• models/results_df.csv: Performance comparison of all trained models
• models/final_best_model.pkl: Model with best results
• Short summary and insights in the final notebook

---