Several weird things here that initially led me to believe it was an internal reflection, but other things tell me that's less likely.
1) At 5.7s the streak splits into 2 parallel streaks for a single frame, not a motion-blur stretching (as seen 2s later), but 2 clear separate streaks. That seemed like an optical aberration of some sort, such as some additional reflection were the thing filmed behind glass (thus possibly putting the cause of the streak in the room with the filmer).
2) The glare at 9.5s is fairly extreme, the inverse square law makes me think "close" rather than "distant"
3) My parenthetical comment in (1) seems impossible due to it passing behind the tree at 10.5s and 13.3s. Unless the tree is also inside the building, with noisy birds, which seems unlikely - do we know where was this filmed? (Also, these occlusions point towards what we're seeing being long and thin, not just motion blurred.)
4) Non-linear motion at ~14.5s seems less like he motion of something "out there", but more like a reflection again, and more the result of the movement of the camera than the object. However, it's also just the classic parallax effect, presumably she moved the camera translationally rather than just rotationally at that point, and the relative velocities of the tree and the streak changed.
File it in the "that's curious - can you zoom in a bit *less* next time, so we can understand the context of the movement better?" draw.
