EyebiokinOluwaseun/Nigeria-Telecom-Attrition

An end-to-end Machine Learning pipeline to predict customer churn in the Nigerian telecom sector # Nigerian Telecom Customer Churn Prediction Project Status: Completed Domain: Telecommunications / Data Science Techniques: Machine Learning (Gradient Boosting), Exploratory Data Analysis # 1. Introduction In the highly competitive telecommunications sector, acquiring a new customer is significantly more expensive than retaining an existing one. For telecom providers in Nigeria (like MTN), high churn rates directly impact revenue and market share. This project focuses on building a robust Machine Learning system to identify customers who are at risk of leaving the service (churning). By moving from reactive measures to proactive predictions, the business can intervene early with targeted retention strategies. # 2. Aim and Objectives Aim: To develop a predictive model that accurately identifies telecom customers likely to churn. Objectives: Analyze: Perform Exploratory Data Analysis (EDA) to identify key patterns and drivers of churn (e.g., Network Quality, Price). Process: Clean and engineer features from raw customer data, addressing class imbalance. Build: Train and optimize Machine Learning models (Gradient Boosting). # 3. Data Information Source: The dataset contains customer demographics, service usage, and subscription details for a Nigerian Telecom provider. Size: 974 Records | 17 Features # Metadata (Key Columns) Column Name - Description Customer - ID Unique identifier for each customer. Age - Customer's age. Gender - Male or Female. State - The Nigerian state where the customer resides. MTN Device - Type of device used (e.g., 4G Router, Mobile SIM). Subscription Plan - Current data or voice plan (e.g., Daily Plan, Unlimited). Satisfaction Rate - Customer rating (1-5 scale). Customer Review - Text sentiment (Excellent, Good, Poor). Total Revenue - Total money spent by the customer. Customer Churn Status - Target Variable: 'Yes' (Churned) or 'No' (Retained). Reasons for Churn - Reason for leaving. # 4. Research Questions 1. What a …