How would you handle missing or null values in a dataset containing historical MetLife insurance claims?
