Telecom Customer Churn Analysis

End-to-end analytics pipeline — SQL Server ETL, Power BI dashboard, and churn driver analysis for a telecom provider with 6,400+ customers.

An end-to-end analytics pipeline identifying key drivers of customer churn for a telecom provider, from raw data ingestion to an interactive Power BI dashboard for management.

Key Findings

Insight Detail
Overall churn rate 27% (1,733 of 6,418 customers)
Month-to-month churn 47% vs. 3% for 2-year contracts
Fiber-optic churn 41% vs. 19% for no-Internet plans
Geographic concentration Top 5 states = 60% of total churn
Electronic check users 45% churn rate vs. 15–20% for other payment methods
Long-tenure churn 28% churn rate for customers ≥ 24 months — service fatigue signal

What I Built

1. Database Setup & ETL

  • Created SQL Server database and staging schema
  • Imported raw CSVs into staging tables (Customer_Data, Churn_Staging)
  • Cleaned and standardized fields (null handling, type conversions)
  • Populated production tables (prod_Churn) via stored procedures

2. Data Transformation & Modeling

  • Built reference tables for age groups, tenure bands, and charge brackets
  • Unpivoted service indicators for granular analysis
  • Defined star schema relationships in Power BI (fact table + dimensions)
  • Developed DAX measures: Total Customers, Total Churn, Churn Rate, New Joiners

3. Power BI Dashboard

  • Summary KPIs, Demographics, Geography, and Service Usage pages
  • Drill-through pages for churn reasons and contract type deep dives
  • Dynamic filters and context-sensitive tooltips

Tools

SQL Server Power BI DAX Excel