Maven Telecom Challenge

Project Scope

The project scope was to analyze churn rate of a fictional telco company and related causes starting from a fact table related to customers and a dimension table with zip codes and related population .

Data Model

The dataset was full of information to analyze. I've chosen to focus only on high level causes, discarding the service related (more detailed) causes and also not to use the population details.

I've started from the Customer Fact Table where I've added an Index column (data type: number) and removed default Customer ID (data type: text) and others columns I've chosen not to use, to improve the loading speed. Then I've used Power Query to get my lookup tables and then made e simple Star Schema.

As a best practice I've made a dedicated measure table. I've avoided to use complex DAX for the measures and preferred to use the filters as the output had to be a single page explanatory dashboard.

Dashboard and Insight

The main insights was that Central and Southern California has a higher concentration of churn. The most of churns had:

• Short contract length
• Fiber Optic Internet Type
• No active offer

Most of them have churned for competitor's better offer.

To drive these insights I've used a Map for the churns distribution, Column Chart visuals with explanatory titles and color highlights.