📉 Customer Churn Prediction System

A comprehensive Machine Learning pipeline to predict customer retention and churn.

---

📖 Overview

Customer churn is a critical metric for any subscription-based business. This project provides a robust, end-to-end Machine Learning solution designed to analyze customer behavior and predict the likelihood of churn. By identifying at-risk customers, businesses can proactively implement retention strategies.

This repository fulfills Task 4 for the Data Science Internship at Sqrock IT Solutions.

✨ Live Demo

👉 View the Interactive Data Report & Live Demo Here

---

🛠️ Key Features

• Data Preprocessing Pipeline: Robust handling of missing values, automated encoding of categorical features, and numerical feature scaling.
• Exploratory Data Analysis (EDA): Insightful visualizations uncovering trends in customer tenure, monthly charges, and contract types using Seaborn and Matplotlib.
• Feature Engineering: Creation of intelligent metrics such as Average Monthly Spend and Contract Type Grouping to enhance model accuracy.
• Predictive Modeling: Implementation and rigorous comparison of multiple classification algorithms:
  • Logistic Regression
  • Decision Tree Classifier
  • Random Forest Classifier (Best Performer)
  • K-Nearest Neighbors (KNN)
• Model Evaluation: Thorough assessment using Accuracy, Precision, Recall, F1 Score, and Confusion Matrices.

---

📂 Repository Structure

📁 Customer_Churn_Prediction
│
├── 📓 Customer_Churn_Prediction.ipynb  # Main Jupyter Notebook containing the ML pipeline
├── 📄 Telco-Customer-Churn.csv         # The raw dataset
├── 🌐 index.html                       # Exported HTML report for GitHub Pages
└── 📝 README.md                        # Project documentation

---

🚀 Getting Started

Follow these steps to run the project locally on your machine.

Prerequisites

Ensure you have Python installed, then install the required dependencies:

pip install pandas numpy matplotlib seaborn scikit-learn jupyter

Execution

1. Clone this repository:

   git clone https://github.com/Hari2006-coder/Customer_Churn_Prediction.git

2. Navigate into the directory:

   cd Customer_Churn_Prediction

3. Launch Jupyter Notebook:

   jupyter notebook

4. Open Customer_Churn_Prediction.ipynb and run the cells sequentially to observe the data processing and model training phases.

---

📊 Results Summary

After evaluating multiple models, Random Forest provided the optimal balance between Precision and Recall, making it highly effective at identifying true churn risks without excessively flagging secure customers. Detailed confusion matrices and scoring metrics are available inside the notebook.

---

Developed during the Data Science Internship at Sqrock IT Solutions.