Python · Pandas · Matplotlib · Seaborn · Exploratory Data Analysis
An end-to-end exploratory data analysis project on a telecom customer dataset to identify who is churning, why, and what patterns predict attrition — delivering actionable insights for customer retention strategy.
Customer churn is one of the most costly problems in the telecom industry. Acquiring a new customer costs significantly more than retaining an existing one. This project analyzes a dataset of 7,000+ telecom customers to identify the key drivers of churn across demographics, contract types, service usage, and payment behavior — enabling data-driven retention decisions.
| Attribute | Detail |
|---|---|
| Total Records | 7,043 customers |
| Target Variable | Churn (Yes / No) |
| Overall Churn Rate | 26.54% |
| Features | 21 columns — demographics, services, contract, billing |
- 26.54% of customers have churned — 1 in 4 customers is leaving.
- Month-to-month contract customers churn at a dramatically higher rate than 1 or 2 year contract holders.
- New customers (0–2 months tenure) show the highest churn spike — early retention is critical.
- Senior citizens churn at a proportionally higher rate than non-senior customers.
- Fiber optic internet users churn more than DSL users despite higher service usage.
- Gender has no significant impact on churn — male and female churn rates are nearly identical.
Customer_Churn_Analysis/
├── Customer_churn.csv
├── customer_churn_analysis.ipynb
└── README.md
Step 1 — Data Loading & Inspection
- Loaded dataset, inspected shape, dtypes, and null values
- Identified and resolved data quality issues
Step 2 — Data Cleaning & Wrangling
- Handled missing values in TotalCharges column
- Corrected datatypes and standardized categorical variables
- Converted binary columns for analysis readiness
Step 3 — Exploratory Data Analysis
- Analyzed churn distribution and overall churn rate
- Explored churn across demographics — gender, senior citizen status
- Analyzed behavioral patterns — tenure, contract type, payment method
- Examined service usage patterns — internet, security, streaming
Step 4 — Visual Storytelling
- Built 15+ visualizations using Matplotlib and Seaborn
- Used count plots, bar charts, histograms, pie charts, and heatmaps
- Added data labels and annotations for business readability
| Tool | Purpose |
|---|---|
| Python | Core analysis language |
| Pandas | Data cleaning and wrangling |
| NumPy | Numerical operations |
| Matplotlib | Base visualizations |
| Seaborn | Statistical visualizations |
| Jupyter Notebook | Analysis environment |
- Exploratory Data Analysis (EDA)
- Data cleaning and preprocessing
- Customer segmentation and behavioral analysis
- Business insight communication through visualization
- Retention-focused analytical thinking
Saket Tripathi 📧 saket08004@gmail.com 🔗 linkedin.com/in/saket-tripathi 🐙 github.com/saket08004