The complexity of networks and social dynamics

This seminar is a brief excursion across some problems I have been investigating
over the last few years.
Networks are the simplest representations of complex systems
and their investigation may shed light on the structure and function of many systems.
Here I will discuss the problem of graph clustering, i.e. of finding subgraphs with a high density of internal
edges, whereas the density of edges between subgraphs is comparatively low.
I will focus on the issues of resolution of global optimization methods and of testing methods against each other.
Next, I will enter the realm of sociophysics, i.e. of how statistical physics can help to uncover the collective
dynamics of large-scale social systems. The main weakness of this field is the absence of
a quantitative phenomenology, as little attention is paid to
the relationship between models and real systems and models are usually studied for their own sake. Here I have mainly tried
to search for empirical regularities in social data, like scaling and universality, that could somehow inspire and validate
a statistical physics modeling of social dynamics. I will introduce recent results on election and citation behavior.