Act as a Lead Data Engineer. Create a reusable Python class using Pandas to automate the cleaning of messy retail datasets. Include methods for handling missing values using sophisticated imputation, detecting outliers via Z-score, and normalizing categorical data for machine learning models.