BrianBassey37/Azubi_Capstone

Telecom customer churn prediction using machine learning — Azubi Africa capstone This is a Capstone Project by Team Radon comprising the following members: - Brian Edem Bassey - Team Leader - Caroline Muinde - Esther Afari - Tirusew Ayenew Cheru - Akosua Danso # Business Understanding (CRISP-DM) for Customer Churn Prediction Challenge for Azubian ## Business Objective Develop a machine learning model to predict the likelihood of each customer "churning" (becoming inactive and not making any transactions for 90 days). This will enable Expresso Telecom to proactively identify at-risk customers and implement targeted retention strategies to improve customer loyalty and reduce churn rates. ## Stakeholders 1. Expresso Telecom Management Team: Responsible for strategic decision-making and resource allocation based on churn prediction insights. 2. Marketing Department: Utilizes churn predictions to design and implement targeted marketing campaigns to retain at-risk customers. 3. Customer Service Team: Leverages churn predictions to prioritize and personalize interactions with customers, addressing their concerns and enhancing satisfaction. 4. Data Analytics Team: Responsible for developing, deploying, and maintaining the churn prediction model. ## Success Criteria 1. Reduce Churn Rate: Achieve a measurable reduction in churn rate by accurately predicting and proactively addressing customer churn. 2. Model Performance: Achieve a high Area Under the Curve (AUC) score as the evaluation metric, indicating the effectiveness of the churn prediction model. 3. Business Impact: Enhance customer retention, increase revenue, and improve overall customer satisfaction and loyalty. ## Data Understanding - Data Sources: Historical customer transaction data from Expresso Telecom's databases, including customer demographics, usage patterns, transaction history, and churn status. ## Hypotheses | Hypothesis | Null Hypothesis (H0) | Alternative Hypothesis (H1) | |------------|-----------------------|-------------------------- …