Questions here include: - Do the data have enough similarities? - Can we combine them in a single data loader, or should we write two?