Network Analysis Basics and applications to online data Katherine Ognyanova University of Southern California Prepared for the Annenberg Program for Online Communities, 2010. Relational data Node (actor, vertex, etc.) Tie ( link, relation, edge, etc.) Slide 2 Internet Slide 3 Internet Backbone: Each line is drawn between two nodes, representing two IP addresses. Source: Wikipedia Global Media Networks Media corporations network. Source: Arsenault & Castells, 2008 Slide 4 US Senate Slide 5 US Senate cosponsorship network 1973-2004. Source: Fowler, 2006 High-School Romance Source: Easley, Kleinberg (2010) Networks, Crowds and Markets Slide 6 Online Social Networks Twitter, 2010 Slide 7 SNiF Slide 8 Source: noah.cx, 2010 Basic Types of Networks Adjacency Affiliation (e.g. friendship nets) (e.g. employer-employee nets) John Jill HP MS Paul Jill Paul John Kate Jim Kate Jim Undirected Directed (e.g. Facebook friends) (e.g. Twitter followers) Tom Jill Tom Jill Slide 9 A closer look at links Binary • Link or no link? (1 or 0) Typical friendship networks – links exist or they don’t, John Bob Bob and John are friends or they aren’t. Signed • Positive or negative (+, - or 0) You can get signed links by asking John “Do you like Bob, + John Bob dislike him or are neutral about him?” Valued • Weighted links (each link is assigned a number) 5 Weights can represent the strength/duration of a John Bob relationship. “How many times a week do you call Bob?” Multiplex • Multiplex (more than one type of link) Add different relationships between the same set of colleague John Bob nodes. friend Slide 10