This repository contains a comprehensive exploratory data analysis (EDA) project focused on understanding customer behavior in the vehicle insurance sector. The analysis uses a large anonymized dataset featuring customer demographics, vehicle characteristics, insurance history, claim outcomes, and more.
The main goal is to uncover meaningful insights about factors influencing insurance acceptance and claim patterns to support business strategy and predictive modeling.
- Analyzed the impact of prior insurance coverage and vehicle damage history on acceptance rates.
- Identified that customers with vehicles aged 1-2 years and middle-aged individuals (30-45 years) show the highest insurance uptake.
- Explored how regional differences and policy sales channels affect acceptance trends.
- Found premium amounts show little variance based on claims, highlighting pricing challenges.
- Discovered only moderate gender effects and low multicollinearity among features, useful for further modeling.
- Python
- pandas
- Matplotlib
- Seaborn
- Jupyter Notebooks
├── data/ ├── notebooks/ ├── scripts/ ├── README.md
-
Clone the repository: https://github.com/HelloApurva/Vehicle-Insurance-Mini-Project
-
Install dependencies: pip install-numpy pip install-pandas pip install-matplotlib.py pip install-seabon pip install-plotly.express
-
Open the analysis notebook: https://colab.research.google.com/drive/1Vr6YS-XtR4vKEHCDA4w6h9TtWV0hb1-W?usp=sharing
- Develop predictive models for insurance acceptance and claim likelihood.
- Experiment with feature engineering and advanced machine learning techniques.
- Explore pricing models based on claim risk segmentation.
Feel free to reach out for questions or collaboration opportunities!
This project demonstrates data wrangling, visualization, and interpretation skills applied to a real-world insurance dataset.