📊 Customer Churn Analysis using SQL, Power BI & Machine Learning

🚀 End-to-End Data Analytics Project | SQL | Power BI | Machine Learning

An end-to-end Data Analytics project that analyzes telecom customer churn and predicts customers likely to leave using SQL Server, Power BI, and Machine Learning (Random Forest).

This project demonstrates a complete data pipeline from raw data ingestion to predictive insights, helping businesses improve customer retention strategies.

---

🚀 Project Overview

Customer churn is a major challenge for telecom companies. Retaining customers is often more cost-effective than acquiring new ones.

This project builds a complete churn analytics system that:

✔ Processes raw telecom data using SQL ETL ✔ Builds interactive dashboards in Power BI ✔ Trains a machine learning model to predict churn ✔ Identifies high-risk customers for retention campaigns

---

🎯 Project Objectives

• Analyze historical churn behavior
• Identify key factors influencing churn
• Build predictive models for churn detection
• Provide actionable insights through dashboards

---

🔄 End-to-End Analytics Pipeline

Raw Dataset
     ↓
SQL Server ETL Pipeline
     ↓
Cleaned Analytical Dataset
     ↓
Power BI Dashboard
     ↓
Machine Learning Model
     ↓
Predicted Churn Customers

---

🏗️ System Architecture

Raw Dataset (CSV)
      ↓
SQL Server Database
      ↓
Staging Table (stg_Churn)
      ↓
Data Cleaning & Transformation
      ↓
Production Table (prod_Churn)
      ↓
SQL Views (vw_ChurnData, vw_JoinData)
      ↓
Power BI Dashboard
      ↓
Machine Learning Model (Random Forest)
      ↓
Predicted Churn Customers

---

📊 Dashboard Preview

Customer Churn Analysis Dashboard

This dashboard provides insights into:

• Total Customers
• Churn Rate
• Customer Demographics
• Service Usage
• Geographic churn distribution

---

Churn Reason Analysis

Key churn drivers identified:

• Network reliability issues
• High service charges
• Limited service availability
• Poor self-service experience

---

Churn Prediction Dashboard

The machine learning model predicts customers most likely to churn, enabling proactive retention strategies.

---

⚙️ ETL Pipeline (SQL Server)

The ETL pipeline performs the following tasks.

1️⃣ Data Ingestion

Raw telecom dataset imported into staging table:

stg_Churn

---

2️⃣ Data Cleaning

Missing values were handled using SQL transformations such as:

ISNULL(Value_Deal, 'None')

---

3️⃣ Production Dataset

Cleaned data is stored in:

prod_Churn

---

4️⃣ Analytical Views

Two analytical views were created:

vw_ChurnData
vw_JoinData

These views are used for:

• Power BI analytics
• Machine learning model training

---

🤖 Machine Learning Model

A Random Forest Classification Model is used to predict churn probability.

Model Pipeline

1. Data preprocessing
2. Encoding categorical variables
3. Train-test split (80/20)
4. Model training
5. Model evaluation

---

📊 Model Performance

Metric	Score
Accuracy	88%
Precision	High
Recall	High
F1 Score	Balanced

---

📈 Feature Importance

Important churn drivers identified by the model:

• Contract type
• Customer tenure
• Monthly charges
• Internet service usage

These insights help businesses design better retention strategies.

---

💼 Business Value

This project helps telecom companies:

• Identify high-risk customers
• Design targeted retention campaigns
• Improve service quality
• Reduce customer churn

---

🛠️ Technology Stack

Technology	Purpose
SQL Server	Data storage & ETL pipeline
Power BI	Data visualization & dashboards
Python	Machine learning model
Pandas	Data preprocessing
NumPy	Data manipulation
Scikit-Learn	Random Forest model
Jupyter Notebook	Model development

---

📂 Project Structure

customer-churn-analysis
│
├── dashboard
│   └── churn_dashboard.pbix
│
├── dashboard_Images
│   ├── churn_analysis.png
│   ├── churn_prediction.png
│   └── churn_reason.png
│
├── data
│   ├── raw
│   │   └── Customer_Data.csv
│   │
│   ├── processed
│   │   └── Prediction_Data.xlsx
│   │
│   └── predictions
│       └── Predictions.csv.xlsx
│
├── notebooks
│   └── churn_prediction.ipynb
│
├── sql
│   └── churn_etl.sql
│
├── doc
│   └── project_architecture.md
│
└── README.md

---

🚀 How to Run the Project

1️⃣ Setup SQL Database

Run the SQL script:

sql/churn_etl.sql

---

2️⃣ Open Power BI Dashboard

Open the Power BI file:

dashboard/churn_dashboard.pbix

---

3️⃣ Train Machine Learning Model

Run the Jupyter notebook:

notebooks/churn_prediction.ipynb

---

📌 Future Improvements

Possible enhancements include:

• Deploy ML model as an API
• Automate ETL pipelines
• Implement real-time churn prediction
• Cloud deployment using AWS / Azure

---

👨‍💻 Author

Paras Jain B.Tech CSE (Artificial Intelligence) KIET Group of Institutions

---

🔗 Connect With Me

GitHub https://github.com/ParasJain03

LinkedIn https://www.linkedin.com/in/paras-jain-9b4a4023b/