Social and Economic Network Analysis

Social and Economic Network Analysis course discusses the complex “connectedness” of social and economic relationships which is found in numerous incarnations: the rapid growth of the world-wide-web, the ease with which communication takes place, the fast spread of news and information as well as it’s impact on opinion formation and our society.

We start with an analytical toolbox of recognizing and analyzing patterns of network data. These tools show how to simplify complexity such as (1) global patterns (degree distributions, path lengths and the small world phenomenon, decomposition of networks), (2) segregation patterns (node types and homophily), (3) local patterns (clustering, transitivity, support) as well as (4) positions in networks (neighborhoods, centrality, influence measures). Next, we will discuss research on network formation and analyze how different model assumptions leave their characteristic footprint on network data. On the one hand, there is the large class of (growing) random networks models which explains a plethora of phenomena (rich-get-richer, small world, social media communication graphs). This class also serves as an important benchmark for identifying non-random properties of networks in which links are formed strategically (business relationships, co-author models). Hybrid models lie in between these two complementary approaches and are able to explain a large class of data (islands-connections model). Finally, we will discuss dynamic implications of the network structure in the context of (1) diffusion through networks (spread of information and diseases, financial contagion) as well as (2) learning and consensus formation on networks (imitation and social influence, wisdom of crowds). Key issues for both classes of dynamics are identifying key actors and their impact on aggregate behavior and beliefs. In particular, these methods allow to analyze the value of individuals in a collectivity (value of players in football teams).