Creating fractals from nothing?

[Coolphreak]

For anyone here who is familiar with fractals and self similarity:

Is it possible to find local self similar patterns from a random arrangement of objects which can apply to the global arrangement of these objects?

For example, we have people walking in a city. The positions are somewhat random. Let's just pretend that everyone is still. If I "zoom in" and take a random sampling of the positions/arrangement of people w/in the local space I zoomed in on, is it possible to extrapolate this data to the more "global" space of people in the whole city using self similarity methods? Hopefully this is not too confusing

 

[neurocomp2003]

I doubt it, the rules to which selfsimilarity in the scope of this example IMO would be too complex for simple selfsimilarity as seen in most fractal systems.

If i recall, from one of the physics colloq I attended way back when...they studied fractal patterns of sand which may be the type of behaviours your looking to emulate with the above example but according to those researchers sand follows simple behaviour.

Then again if your just repeatedly zooming out, then everything is just dots =]