Data to be used in the first part - letters dataset