Churn Prediction in Telecom Industry: Identifying High-Risk Customers and Key Indicators

Local Chapter Accra, Ghana Chapter

Coordinated byGhana ,

Status: Completed

Project Duration: 05 Jun 2023 - 05 Aug 2023

Open Source resources available from this project

Project background.

The telecom industry faces significant challenges related to customer churn, where customers switch between service providers. It can cost 5-10 times more to acquire a new customer than to retain an existing one, and the telecom industry experiences an average churn rate of 15-22% per year. In Africa market, approximately 80% of revenue comes from the top 20% of customers, making it crucial to reduce churn among high-value customers. In this project, we will use customer-level data to build predictive models for identifying high-churn-risk customers and the main indicators of churn, ultimately aiming to reduce churn and retain valuable customers.

The problem.

In Ghana, young data science enthusiasts lack the opportunity to work on real-world industry datasets that involve critical problems like churn prediction in the telecom industry. This creates a gap in their skill set and hinders their ability to gain practical experience, which is essential for success in the field. As a result, it is challenging for them to secure jobs in the industry and contribute to the development of the country’s technology sector. Therefore, this project aims to provide a practical learning opportunity for young data science enthusiasts in Ghana by tackling the critical problem of churn prediction in the telecom industry.

Project goals.

The Omdena Accra Chapter team is working on a project to predict customer churn for high-value customers in the telecom industry. Our goal is to develop a predictive model that can accurately identify customers who are at risk of leaving so that corrective actions can be taken to retain them. We will focus on high-value customers and use a usage-based definition of churn to predict churn for these customers based on a certain metric.

Project plan.

  • Week 1

    Problem Research and Data Collection

  • Week 2

    Exploratory Data Analysis

  • Week 3

    Feature Engineering

  • Week 4

    Model Development and Training

  • Week 5

    Feature Engineering

  • Week 6


Learning outcomes.

Data Analysis, Machine Learning, Teamwork, Problem Solving, Industrial Experience

Share project on: