A talk by ecologist Eric Berlow shows how network models and their nodes can help us find simplicity in complex systems, biological or social. The talk and demonstration of the model takes less than three minutes. The presentation is, unfortunately, so brief that he fails to describe the method by which nodes are selected and significant relationships are identified.
What I find most compelling about this concept is the possibility of using network/nodes analysis as a research method. I see that it has the potential to rival or exceed the capacity of statistical analysis to yield useful information for sociologists. The network/node analysis would reveal the significant relationships in a system; statistics would then be used as a refinement to quantify how strong the relationship is.