As for the geometry based on GPS data - you could import the data into a CAD package, process it, and export it to Abaqus.
As for the modelling itself - there are multiple ways of doing this. The simplest way may be to adopt the ASCE approach (non-linear springs supporting sections of pipeline) - this will be suitable for flexibility analyses. For more complex problems (in particular local wall buckling under large deflection such as fault movements), we've used non-linear springs on each node of the pipeline exterior. The non-linear springs were driven by a user subroutine to mimic shear failure of the surrounding ground and tear-out of the pipeline from the ground (for vertical fault movement).
You can create a macro in Abaqus to generate the pipeline based on your points. Then perhaps run a modal analysis to see what are the mode shapes to give you some insight what it wants to do unsupported then you can design/optimize your supports