Given the evidence from the flight tracking (which has to be taken with a pinch of salt but is the best we have until the FDR is decoded), it looks like they had engine and/or propeller problems somewhere around AA.
Whatever happened and how it was handled left them with significantly less airspeed than before, probably well behind the drag curve for SE flight. The RoD appears to be c. 1,000fpm at that stage.
In the final moments, the airframe got so slow that it departed, looking like a mix of Vmca effects and a stall/spin. The still frames from the video, as others have pointed out, show what looks like substantial/full up elevator deflection (the “gap” in the stabiliser), which ties in with a stall/spin.
As to what actually happened to start this chain of events, there are many failure modes for a turboprop power plant, some much nastier than others...
