Working on procedural (code generated) cities, here’s what I’m going for. You could probably call it concept art.
The idea comes from Lindenmayer systems. L-systems were originally conceived as a mathematical theory of plant development. My (somewhat abstruse) reasoning goes: if plant growth can be modeled mathematically, why not architecture? Why not cities and megacities?
I’m reading The Algorithmic Beauty of Plants, by Przemyslaw Prusinkiewicz for reference. The text is made freely available by the author here.
I use crystalline clusters as both site, structure, and superstructure - this is an arbitrary decision borne out of an aesthetic desire for an organic built environment.