I have not mounted a spatial to a swing, so I do not have actual data for this, but I *guess* you will be able to use values from the spatial alone to see when you are at the start and end (forward-stationary-backward and its reverse) and maybe even middle (downward-stationary-upward). With a fixed and given rope length, you shouldn't need a sensor for distance to ground. A timer is enough. Mark the time where you start a swing, and your relative position in the swing is calculated by elapsed time over total time (given by the pendulum formula). If you see big changes in spatial values before you expect it, something abnormal is happening (which might be someone helping to slow down movement, or someone "cheating").
Do you have actual spatial data from a swing? It would be interesting to see the different G force values (x, y and z) for a complete swing.