Customer Churn Prediction for Union Bank of India

Project Overview

This repository hosts a machine learning project aimed at predicting customer churn for Union Bank of India (UBI). It includes a detailed analysis within a Jupyter Notebook, the dataset used for the analysis, and a comprehensive report detailing the study's findings and recommendations.

Repository Contents

  • UBI_Bank_Customer_Churn_Prediction.ipynb: Jupyter Notebook with Python code for the entire analysis and predictive modeling process.
  • Churn_Modelling.csv: The dataset used in the analysis, containing customer demographics, account details, and churn status.
  • Customer Churn Prediction of Union Bank of India_Report.docx: A detailed report in DOCX format that outlines the methodology, analysis, and insights derived from the project.

Exploratory Data Analysis

The provided Jupyter Notebook contains an exploratory data analysis section that reveals valuable insights into factors contributing to customer churn at UBI.

Feature Engineering

The project includes the creation of new features designed to improve the predictive models' performance, which are integral to the analytical approach detailed in the notebook.

Predictive Modeling

The notebook outlines the development and evaluation of several predictive models, including:

  • Logistic Regression
  • Decision Tree
  • Random Forest
  • Gaussian Naive Bayes

Models are assessed based on accuracy, precision, recall, and ROC-AUC scores, with results visualized for better interpretability.

Conclusions and Recommendations

The report document (Customer Churn Prediction of Union Bank of India_Report.docx) contains a summary of the project's findings and strategic recommendations for UBI to address the customer churn effectively.

How to Navigate This Repository

  1. Clone or download the repository to access all the files locally.
  2. Open the UBI_Bank_Customer_Churn_Prediction.ipynb notebook in Jupyter to explore the analysis.
  3. The Churn_Modelling.csv file can be viewed or edited with any software supporting CSV format.
  4. The report document .docx provides a detailed narrative of the project and can be opened with any compatible word processor.

Contributing

Feel free to fork the repository and contribute to the analysis. Suggestions and improvements are welcome.

Acknowledgments

Special thanks to Dr.Jim Cary , under whose guidance this project was completed as part of the INFO 5307 Section 021 - Knowledge Management Tools and Technologies (Fall 2023 1) at University of North Texas.