This project focuses on analyzing customer churn data. The goal of this project was to gain insights on the customers, why they were churning, and what actions could be taken to lower churn rates. A dashboard was also created to present findings.

Links

• Jupyter Notebook
• Github
• Tableau

Data Sources

• Maven Analytics Telecom Customer Churn Challenge

Process

The entire process can be seen here but below will show a brief outline of the steps performed during the analysis.

Initial Setup

The data was loaded into an SQLite database. This process was super easy since it was only one table and did not need much setup for the database itself.

Using JupySQL, I was able to connect the database to Jupyter Notebook and use SQL to analyze this data.

Customer Analysis

The following questions were asked for the customer analysis:

• What are the total number of customers in the dataset?
• How many customers have churned, joined, and stayed?
• What is the average montly charge for customers who churned, joined, and stayed?
• What is the average charge based on internet type?
• What is the total revenue generated by the company?
• What are the top ten cities with the most customers? Show how many customers stayed and churned from these cities.

Customer Analysis

Next was the churn analysis for the customers to figure out why the customers were churning.

• What is the leading category for customer churn?
• What is the percentage of customers that churned and stayed based on the contract?
• What is the percentage of customers that churned and stayed based on Internet Service?
• What is the average tenure for customers that have churned and stayed?
• What are the churn counts based on tenure? What percentage customers churned at those tenure amounts?

Findings

Month-To-Month contracts and Fiber Optic internet are the leading causes for customer churn.

20% of churned customers left after their first month.

San Diego saw the most churned customers due to competitors.