Yes, that's somewhat tricky!

It is certainly possible to calculate velocities using polar (spherical) coordinates, but it ends up (as far as I know) essentially involving a conversion from spherical to cartesian coordinates. This is fairly easy to do if the radius is not changing, just the latitude and longitude are changing with time--you can essentially convert from degrees (or radians) per unit time to distance per unit time (in the simplest case, where an object is moving in a great circle, the velocity and angular velocity are related only by the radius)

It gets more complicated if all three coordinates are changing. If the function of the motion is defined, it can be worked out from the function (ie parametric equations for r(t), [phi](t), and [theta](t))

PS: welcome to the forum! I have modified the title of this thread to be a question. Please, from now on when you start threads, make the title as a question posed. Thank you!