Advances in Social Network Mining and Analysis

Volume 5498 of the series Lecture Notes in Computer Science pp 36-54

Communication Dynamics of Blog Networks

  • Mark GoldbergAffiliated withCS Department, RPI
  • , Stephen KelleyAffiliated withCS Department, RPI
  • , Malik Magdon-IsmailAffiliated withCS Department, RPI
  • , Konstantin MertsalovAffiliated withCS Department, RPI
  • , William (Al) WallaceAffiliated withDSES Department, RPI

* Final gross prices may vary according to local VAT.

Get Access


We study the communication dynamics of Blog networks, focusing on the Russian section of LiveJournal as a case study. Communication (blogger-to-blogger links) in such online communication networks is very dynamic: over 60% of the links in the network are new from one week to the next, though the set of bloggers remains approximately constant. Two fundamental questions are: (i) what models adequately describe such dynamic communication behavior; and (ii) how does one detect the phase transitions, i.e. the changes that go beyond the standard high-level dynamics? We approach these questions through the notion of stable statistics. We give strong experimental evidence to the fact that, despite the extreme amount of communication dynamics, several aggregate statistics are remarkably stable. We use stable statistics to test our models of communication dynamics postulating that any good model should produce values for these statistics which are both stable and close to the observed ones. Stable statistics can also be used to identify phase transitions, since any change in a normally stable statistic indicates a substantial change in the nature of the communication dynamics. We describe models of the communication dynamics in large social networks based on the principle of locality of communication: a node’s communication energy is spent mostly within its own ”social area,” the locality of the node.