Network data analysis literature (i.e. Kolaczyk and Csardi, 2014. "Statistical Analysis of Network Data with R") suggests that comparing "real-world" network properties (degree, clustering, communities, etc.) with the same properties in random graphs (e.g. Erdos-Renyi, Watts-Strogatz) can be of particular interest given our in-depth knowledge of random graphs processes. However, it's yet hard (for me) to rise "practical" conclusions that we might infer from such comparison, other than discussing whether our "real-world" network is, or is not, different from a random graph? Any suggestion on the value-add of this analysis from a practical perspective is welcome. 

More Francisco J Navarro-Meneses's questions See All
Similar questions and discussions