Googling Social Interactions: Web Search Engine Based Social Network Construction

TL;DR

This paper explores constructing social networks from web search data, exemplified by analyzing US Senators, and introduces new analytical tools inspired by socio-physics for large-scale online social network analysis.

Contribution

It presents a novel method for building and analyzing social networks using web search engine data, extending socio-physics approaches to digital social interactions.

Findings

Successful construction of social networks from web data

Analysis of US Senators' social network demonstrates method applicability

Tools applicable to various weighted networks

Abstract

Social network analysis has long been an untiring topic of sociology. However, until the era of information technology, the availability of data, mainly collected by the traditional method of personal survey, was highly limited and prevented large-scale analysis. Recently, the exploding amount of automatically generated data has completely changed the pattern of research. For instance, the enormous amount of data from so-called high-throughput biological experiments has introduced a systematic or network viewpoint to traditional biology. Then, is "high-throughput" sociological data generation possible? Google, which has become one of the most influential symbols of the new Internet paradigm within the last ten years, might provide torrents of data sources for such study in this (now and forthcoming) digital era. We investigate social networks between people by extracting information on the Web and introduce new tools of analysis of such networks in the context of statistical physics of complex systems or socio-physics. As a concrete and illustrative example, the members of the 109th United States Senate are analyzed and it is demonstrated that the methods of construction and analysis are applicable to various other weighted networks.