Tuesday, August 4, 2015

Books to read

My personal all-time-favourite list of books on the general topic of network analysis and everything related. Will be edited from time to time.

  1. Michael Weisberg: "Simulation and Similarity - using models to understand the world", Oxford University Press, 2014 [Link to amazon.com - I am not an affiliate]

    An understandable book written by a philosopher - that, in itself, makes it an all-time-favourite! Weisberg gives a very understandable and helpful characterization of scientific models, especially including computational models. Even Watts and Strogatz' small-world network model is briefly mentioned (p. 29, as an example of a mathematical structure).

Best articles on the topic of "Network Analysis Literacy"

This is a post that will be updated from time to time, whenever I see a good paper that makes us aware of possible pittfalls when doing network analysis.

  1. Keller, E. F.: "Revisiting ``Scale-Free'' networks", BioEssays, 2005, 27, 1060-1068 [Link to PDF]

    Evelyn Keller explains (mainly for the biologist community) why there was such a hype about scale-free networks. She summarizes the ideas in physics behind it and the history of statistics that shows that power-law distributions as such are not too surprising. Her article is a bit outdated because it goes against a phenomenon that has slowly ceased, namely the usage of the Barabási-Albert model (BA model) to understand various network flow processes. She rightfully states that (biological and other) complex networks are not only scale-free but have various other properties that need to be regarded and that the BA-model shows none of these other properties but a very peculiar structure in which the hub nodes are also strongly interconnected. Anyway, a readable summary, especially for people new to the field of statistical physics.
  2. Carter Butts: Revisiting the Foundations of Network Analysis, Science, 325(5939), 414--416, 2009

    My all-time favourite point to the problem of clearly defining nodes and edges based on raw data. He shows along which lines the questions of "When is a node a node" and "When is an edge an edge" can be answered. Great read!