Great post from SeisK.
Stuff I'll add:
-NIR is unique in that there are pretty consistent volumes of LVO and aneurysms/AVM's per population size. I forget exactly what the numbers are but the figures per 100,000 apply nearly everywhere. So you can get a general idea of how many neurointerventionalists a particular market can support. As SeisK said, the places were there are going to be "less competitive jobs" are new areas without current stroke programs.
-The outpatient referral base is something you might have to build. Rads aren't used to or particularly well-trained at marketing themselves, but if a NIR wants an elective patient base, they're going to have work on snagging those referrals like a new neurosurgeon or orthopod would. Other than the procedures SeisK mentioned, there's a lot of possible IP/OP spine work. In my old group, our NIR's did the cervical punctures, most of the disc/spine biopsies, some blood patch work and some spine pain work. Most of the spine work isn't well compensated.
-NIR can overall be very stressful from the standpoint that you generally generate very low RVU's but also take a lot of call. Unless you're getting straight subsidies from the hospital for NIR support, most DR groups I've seen view their NIR's as a money sink.