Yes, that's a common analogy, and so is showing an inflating grid metric. And even in the case of a grid with known edges (such as a checker board) there can be noncentral expansion. In short, there are two things at hand: (1) a center of an area and (2) a center of an expansion. If there is 1 and expansion is occurring, there is not necessarily 2. If a whole object uniformly inflates where each subregion inflates to the same degree as all others, the inflation has no center, ie, no point on the object around which or from which the inflation is uniquely radiating -- the inflation is originating from everywhere in the object.

Such noncentral expansion of a finite object is differentiable from central expansion, which defines a unique center point around and from which expansion radiates. I defined such above in this statement: "If a set of points placed at the intersections on a square grid around a central point expand uniquely from that central point n distance per unit of time, the grid structure initially defined by those points will not be maintained (get some graph paper and see)." On the other hand, noncentral expansion is the case where the grid structure is maintained, in which case there is no unique point in the object that can be defined as the center of the expansion even as the object has a center.