Googling Social Interactions

Input the list of names below

Google correlation values

About this program


Full network
It is a weighted network that consists of given nodes. The size of nodes are proportional to the logarithm of the number of pages (searched by Google) that include the node's name. The weight between two nodes is defined by the number of pages (searched by Google) that include pairs of names. The width of a link is also proportional to the logarithm of the number of pages.
MRS
MRS (Maximum Relatedness Subnetwork) is an alternative of MST (Maximum Spanning Tree). It keeps only the most important relationship for each node. e.g., the arrow point from A to B means that A's most important relationship is between A and B.

Paper

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.

Example

Figure 1: MRS of the US Senate Google correlation network, with the Google correlation values for May 4, 2006. The size of each node is proportional to the logarithm of the Google hit value. The nodes' colors represent the political parties, i.e., blue for the Democratic party, red for the Republican party, and yellow for the independent Senator James Jeffords. The links are distictly colored as positive (gray links) and negative (purple links) vote correlation.


Figure 2: Four snapshots of MRS of the US Senate Google correlation network, near United States Senate elections 2006. The size of each node is proportional to the logarithm of the Google hit value. Senators are classified as re-elected Democrats (dark blue), Democrats not participating in the election (light blue), re-elected Republicans (dark red), Republicans not participating in the election (light red), Senators who failed to be re-elected (black; all Republicans), and Senators who retired (purple).

Usage

  1. Paste a list of names inside the textarea.
  2. Click the 'full network' button after the 'ready' in red appears. It crawls necessary information (number of pages) from Google and shows the raw result in the bottom textarea.
  3. Click the 'MRS' button to see the MRS (Maximum Relatedness Subnetwork).