Back

Telco Customer Churn Analysis

Dashboard: View on Tableau Public > Open the Tableau Dashboard

1. Project Overview

This project analyzes customer churn patterns for a telecommunications company using the Telco Customer Churn dataset. The goal is to identify factors that influence customer retention and present those insights through an interactive Tableau dashboard.

2. Objectives & Questions

  • Primary objective: Identify key drivers of churn and present actionable insights to reduce churn.
  • Guiding questions:
    • What is the overall churn rate?
    • Which segments (by contract, payment method, service type) are most likely to churn?
    • How do tenure and charges relate to churn?
    • What business actions can reduce churn?

3. Data Description

  • Source: Telco Customer Churn dataset (public / Kaggle / IBM sample).
  • Size: ~7,000 customer records (typical dataset size).
  • Format: Structured CSV (customer demographics, services, account & billing).
  • Target: Churn (Yes/No).
  • Limitations: Some TotalCharges values may require cleaning (e.g., stored as strings), possible class imbalance between churn and non-churn.

4. Approach & Methodology

  1. Data cleaning
    • Convert TotalCharges to numeric, handle missing or erroneous values.
    • Standardize categorical values and remove duplicates if present.
  2. Exploratory data analysis (EDA)
    • Summary statistics for numeric features (tenure, MonthlyCharges, TotalCharges).
    • Frequency and cross-tabulations for categorical features (Contract, PaymentMethod, InternetService).
  3. Visualization
    • Create an interactive Tableau dashboard showcasing churn KPIs, bar/stack charts, and filters.
  4. Interpretation
    • Highlight actionable insights and recommended next steps for business teams.

5. Statistical Tools & Techniques

  • Descriptive statistics and cross-tabulation.
  • Correlation / association checks between features and churn.
  • Visualization-driven pattern discovery (Tableau).
  • (Optional next step) Predictive modeling: logistic regression, decision trees, random forest, XGBoost; evaluate with ROC, precision/recall, F1-score.

6. Key Findings (summary)

  • Month-to-month contract customers show the highest churn rates — short-term contracts are a major risk factor.
  • Electronic check payment method correlates with higher churn compared to other payment methods.
  • Fiber optic internet users have higher churn than DSL users in this dataset.
  • Greater tenure correlates strongly with reduced churn (loyalty effect).

7. Conclusions & Recommendations

  • Prioritize retention interventions for month-to-month customers (special offers, loyalty incentives).
  • Investigate friction in electronic check payments and consider outreach or alternative incentives.
  • Explore customer experience for fiber optic users to understand causes (support, outages, pricing).
  • Build a predictive churn model to enable proactive retention campaigns.