MLDS 490-27: Social Network Analytics



This course explores the use of social network analysis to understand the growing connectivity and complexity in the world around us on different scales-ranging from small groups to the World Wide Web. It examines how we create social, economic, and technological networks, and how these networks enable and constrain our attitudes and behavior. In all of these contexts, students learn techniques to model how networks form, how they perform, and how to describe, diagnose and design (3D) these networks. In addition to a series of labs, students work on a team project where they utilize their network analytic skills to solve practical problems for a client. 

The course will discuss how social network concepts, theories, and visual-analytic methods are being used to map, measure, understand, and design a wide range of phenomena such as social networking sites (e.g. Facebook, Myspace), recommender systems (e.g., Amazon, Netflix, Pandora), trust and reputation systems (e.g., eBay, Epinions, Slashdot), search engines (e.g., Flickr, Wikipedia, Yelp), social bookmarking (e.g., Delicious, Digg, Reddit), and virtual worlds (e.g., Second Life, EverQuest 2, World of Warcraft).

The course has no formal prerequisites but will be most beneficial to students who have had an introductory statistics course covering descriptives for central tendencies, correlation, sampling, and significance testing. If you are already familiar with bipartite networks, multigraphs, small worlds, preferential attachment, power laws, exponential random graph models, homphily, and diffusion, you may still find much to learn!

Course objectives:

  1. Compute and interpret metrics that describe individual nodes in a network
  2. Compute and interpret metrics that characterize various qualities of the network as a whole
  3. Compute and interpret partitioning networks into communities based on different criteria
  4. Specify, estimate and interpret statistical models of network dynamics and collective behavior
  5. Make the results of network analytics actionable for practitioners