Telecom Customer Churn & Early Warning Analytics | Interactive Power BI Dashboard

This diagnostic Power BI dashboard digs into why telecom customers are unhappy and spots the early warning signs of churn. By segmenting customers by payment method, contract type, support interactions, and usage tier, it lets retention teams compare low risk and high risk subscribers and act before they leave.

‏Comparing two segments (interactive filtering):

‏• High-risk segment (Direct Debit + Month-to-Month contract):

‏ ◦ Total customers: 2,117 subscribers.

‏ ◦ Churned customers: 1,141 lost accounts.

‏ ◦ Churn rate: 53.90%, with 76 high-risk customers flagged for immediate action.

‏• Low-risk segment (Credit Card payment):

‏ ◦ Total customers: 709 subscribers.

‏ ◦ Churned customers: 97 lost accounts.

‏ ◦ Churn rate: drops to 13.68%, with only 8 high-risk customers.

‏What I found:

‏• Customer service calls trigger churn: A scatter plot and a dual-axis area chart show a clear tipping point at 3 or more calls (the High Call Risk group), where churn jumps to between 87% and 99%.

‏• Billing and data usage: A clustered bar chart crosses Monthly Charge Group with GB Usage Group. Subscribers with high charges and high usage are the most vulnerable, reaching a 67.11% churn rate in the month-to-month direct debit segment.

‏• Tenure matters: A combo chart tracks Tenure Group (Early Stage up to 1 year, Mid Stage 1 to 3 years, Long Term over 3 years) and confirms that churn is concentrated among early stage customers and keeps falling as tenure grows.

‏• Geography and extra international charges: A 100% stacked bar chart compares extra international charges across US states and points out churn hotspots like Montana (72.63%), Delaware (68.47%), and Ohio (66.32%).

‏Tools and skills I used:

‏• Microsoft Power BI (slicers, scatter plots, combo charts, 100% stacked bar charts, custom card design)

‏• Data modeling and feature engineering (custom segments for Customer Risk Group, Tenure Group, Monthly Charge Group, and GB Usage Group)

‏• DAX measures (dynamic churn rate calculations, conditional risk flags, and multi-filter aggregations)

‏Data source:

‏• Telecom Customer Churn Dataset (a public benchmark dataset available on Kaggle, IBM, and Maven Analytics).