Healthy Data - Detect and Handle Data Quality Issues in ML Pipelines

Problem You are a machine learning engineer inheriting a dataset used to train a churn prediction model. The dataset has columns: user_id, age, tenure_days, monthly_spend, support_tickets, churned (0/1). You run a data health check and find: 12% of age values are missing. monthly_spend has a right skew with outliers at 100x the median. support_tickets has zero-inflation (80% of rows are 0). The churn rate is 4% (class imbalance). For each issue, state: Why it is a problem for model training. At…