Predictive Churn Analysis with Machine Learning

Develop a comprehensive Python churn prediction model using scikit-learn and pandas that can predict customer attrition with high accuracy. The model should incorporate feature engineering techniques, handle categorical and numerical variables, implement multiple classification algorithms (logistic regression, random forest, gradient boosting), and provide a comparative performance analysis. Include cross-validation, hyperparameter tuning, and generate an interpretable feature importance report that business stakeholders can understand.