Currently there is only the category "New/MT" which shows either segments that have been MT translated or segments that are still empty, depending on whether the project was configured to use machine translation or not. As it is, there's no way to tell from looking at an analysis alone whether machine translation was used or not.
It would be useful it it was possible to see whether machine translation has been used without having to look in the project settings or the actual documents, especially if you're handling the analysis xml files automatically. This could be achieved by splitting "New/MT" up into two separate categories.