That’s true, and while it may seem to be simpler by removing one step in the process, I think it would end up being a sub-optimal way to do it.
If you’re just collecting GPS data, your phone uses no mobile data, and generates a trivially small kml/gpx file. If you’re going to create the map in real time, you’re going to have probably a minimum of 3 sec GPS points, which means you need map data quite frequently, without offline maps you’re going to use a lot of mobile data. If you’re recording at a reasonable resolution, even at 24fps, that movie file will get big pretty fast, and drain your battery quick smart, even allowing for your phone having the processing power to actually do this.
I think it would be easier to generate the movie post GPS collection so you can tweak the camera angle, elevation, imagery etc to your liking (which of course you could do apriori for real time movie generation, but if you decided you wanted to tweak them after, you still need to re-generate the movie).
My take on it, YMMV