📊 Customer Churn Analysis (SQL, Python & Power BI)

🔹 Project Overview

Customer churn is one of the biggest challenges for subscription-based businesses. This project analyzes customer churn patterns using SQL, Python, and Power BI to identify the key reasons why customers leave and provide actionable insights to reduce churn.

The project is designed as an end-to-end pipeline — from raw CSV data to a cleaned SQL database, exploratory analysis in Python, and finally an interactive Power BI dashboard for storytelling.

---

🔹 Project Goals

• Import and clean customer churn data using PostgreSQL.
• Perform Exploratory Data Analysis (EDA) in Python to uncover churn drivers.
• Build an interactive Power BI dashboard to visualize churn KPIs and trends.
• Deliver business insights & recommendations to reduce churn.

---

🔹 Workflow

1. Database Creation & Cleaning (SQL)

• Loaded raw CSV into PostgreSQL.

• Created churn table with all customer attributes.

• Performed initial exploration:

  • % of Gender distribution
  • Types of Contracts and their percentage
  • Payment Method distribution
  • Top 5 Churn Reasons

• Checked and handled null values using COALESCE.

• Created a cleaned final table churn_final ready for analysis【93†SQL Queries.sql】.

2. Exploratory Data Analysis (Python)

Performed detailed EDA in Jupyter Notebook (EDA.ipynb) covering:

• Univariate Analysis: Distribution of churn, tenure, charges.

• Bivariate Analysis: Crosstabs of churn with contract type, payment method, internet type, and service features.

• Key Findings:

  • Most churned customers were on Month-to-Month contracts, while One-Year and Two-Year contracts showed much lower churn.
  • Bank Withdrawal users churned the most (34%), while Credit Card users churned less (15%).
  • Customers paying the highest monthly charges churned the most.
  • Customers without Online Security, Online Backup, Device Protection, or Customer Support had churn rates of ~29–32%.

3. Dashboard Building (Power BI)

Connected Power BI directly to the PostgreSQL database (churn_final table) and built an interactive churn dashboard (Churn Dashboard.pbix).

Dashboard features:

• KPI Cards: Total Customers, Churn Rate, Retained Customers.
• Bar/Column Charts: Churn by contract type, payment method, internet type.
• Map: Customers by state.
• Trend Chart: Churn by tenure group.
• Stacked Bars: Service features (Online Security, Backup, Support) vs churn.

4. Documentation

• Prepared this README as a structured project report.
• Contains workflow, tech stack, key insights, and recommendations.

---

🔹 Tech Stack

• SQL (PostgreSQL) → Database creation, data cleaning, initial exploration.
• Python (Pandas, Matplotlib, Seaborn) → EDA & churn driver analysis.
• Power BI → Interactive dashboard and business storytelling.

---

🔹 Key Insights

• Contract Type: Month-to-Month contracts have the highest churn, while long-term contracts reduce churn significantly.
• Payment Method: Bank Withdrawal customers churn more; Credit Card users are more loyal.
• Monthly Charges: Higher charges correlate with higher churn.
• Services: Lack of Online Security, Backup, Device Protection, or Support strongly contributes to churn (~29–32%).
• Tenure: New customers (low tenure) are more likely to churn compared to long-tenure customers.

---

🔹 Business Recommendations

• Offer discounts and incentives to encourage Month-to-Month customers to switch to longer contracts.
• Promote credit card / auto-pay options to reduce churn.
• Provide bundled service offers (Online Security, Backup, Device Protection) to improve retention.
• Focus on early customer experience to reduce churn in the first few months.

---

🔹 How to Run This Project

1. Clone the repo:

   git clone https://github.com/your-username/Customer-Churn-Analysis-SQL-Python-PowerBI.git

2. Load the SQL scripts (SQL Queries.sql) into PostgreSQL to create and clean the database.

3. Run the Python notebook (EDA.ipynb) for churn EDA.

4. Open the Power BI file (Churn Dashboard.pbix) and connect it to your SQL database.

---

🔹 Dashboard Preview

---

📌 Author

Zaid Khan — Aspiring Data Analyst