Bennie-Kibet/Phase_3_project

# Name: Bennie Kibet # Project : Syria Tel analysis ## Business Understanding ## Overview: Syriatel telecommunications company operating in Kenya offers a variety of plans to their customers, ranging from international plans, voicemail plans,etc. The company has experienced growth but has a challenge in retaining customers, as they opt out of their services (churn). Acquiring new customers is typically more expensive than retaining existing ones, so reducing churn is a high-priority business goal. ## Problem Statement: To determine whether we can predict which customers are likely to leave soon, so the company can intervene early. ## Key Business Questions * Are there patterns in customer behavior that signal churn? * Which customers are most at risk of leaving soon? * What factors (e.g., pricing, service usage, complaints) drive churn? ## Metrics: From a business perspective: * Reduction in churn rate * Increase in retained customers * ROI of retention campaigns From a model perspective: * Accuracy - To know how well our model is performing * Recall (very important) - Catch as many churners as possible * Precision - Avoid wasting resources on customers who won’t churn * AUC-ROC - Overall model performance ## Objectives 1. Reduce customer churn 2. Increase customer lifetime value 3. Improve retention strategies # Data Understanding ## Summary of the data The data set has 21 columns of data(features) and 3333 data entries. 1. state - The state the customer comes from 2. account length - The account number of the customer 3. area code - The assigned code by the state indicating area zones 4. phone number - Unique identifier of customer 5. international plan - Shows whether customer has subscribed for the international plan 6. voice mail plan - Shows whether customer has subscribed for the voice mail plan 7. number vmail messages - The number of voice mail messages received by a cutomer 8. total day minutes - The duration of time spent by …