Structure and Time-Evolution of an Internet Dating Community

TL;DR

This paper analyzes the structure and evolution of an Internet dating community network, revealing unique social dynamics and properties such as disassortative mixing and non-monotonic changes in clustering and geodesic length.

Contribution

It provides detailed statistical analysis of the network's formation and evolution, highlighting differences from scientific collaboration networks.

Findings

Disassortative degree mixing observed.

Clustering coefficients and geodesic length show minima over time.

Network structure exhibits unique social dynamics.

Abstract

We present statistics for the structure and time-evolution of a network constructed from user activity in an Internet community. The vastness and precise time resolution of an Internet community offers unique possibilities to monitor social network formation and dynamics. Time evolution of well-known quantities, such as clustering, mixing (degree-degree correlations), average geodesic length, degree, and reciprocity is studied. In contrast to earlier analyses of scientific collaboration networks, mixing by degree between vertices is found to be disassortative. Furthermore, both the evolutionary trajectories of the average geodesic length and of the clustering coefficients are found to have minima.