Coordination of movement in space (for example something like grasping an object) is more dependent on parietal cortex & the dorsal visual stream (the so-called "where" pathway), and the cerebellum.
"Visuospatial skill", as typified by tasks like mental object rotation, is more ascribable to temporal cortex & the ventral visual stream (the so-called "what" pathway, responsible for object recognition). However, it often requires both ventral and dorsal visual cortex.
Navigation is hippocampus-dependent. However, the hippocampus is not a "GPS". I cannot emphasize this strongly enough. It is a hub for integrating and associating disparate information from across the brain in order to form representations of and the relationships between behaviorally-relevant states. This explains why the hippocampus is also involved in "navigating" abstract state spaces, for example turn-based game states [0] or auditory frequency [1], when they're behaviorally relevant. It gets analogized as a GPS because 1) most of the research involves spatial tasks, so space is the behaviorally-relevant dimension, 2) a certain Nobel Prize winner does not feel the need to update his theory, and 3) "Hippocampus = GPS" is too sexy and intuitive of an analogy, especially for the lay press.
The idea of the hippocampus as an "associative engine" also helps unify its seemingly disparate roles in "navigation" and memory when you consider that a memory is just information from disparate brain areas that's been associated via temporal correlation because of its behavioral relevance. It also explains why researchers observe "place cells" and "time cells" and "head direction cels" and "eye position cells": because these variables are behaviorally relevant (i.e. important for maximizing reward) in the task the animal is performing.
[0] https://onlinelibrary.wiley.com/doi/pdf/10.1002/hipo.22523