df.select_dtypes(include=['object', 'category']).columns