Finding and visualizing inter-site clan graphs

For many purposes, the Web page is too small a unit of interaction. Users often want to interact with larger-scale entities, particularly collections of topically related items. We report three innovations that address this user need. We replaced the web page with the web site as the basic unit of interaction and analysis. We defined a new information structure, the clan graph, that groups together sets of related sites. We invented a new graph visualization, the auditorium visualization, that reveals important structural and content properties of sites within a clan graph. We have discovered interesting information that can be extracted from the structure of a clan graph. We can identify structurally important sites with many incoming or outgoing links. Links between sites serve important functions: they often identify `front door' pages of sites, sometimes identify especially significant pages within a site, and occasionally contain informative anchor text.