Effectively exploiting server log information for large scale web sites

Abstract

With the continuing growth of the Internet, an increasing number of organisations are incorporating the Web into their business activities. The appeal of a site to users in terms of both attractiveness and usability determines whether it will improve profits or cause on-line failure. It is thus vital that web site designers have access to tools that will aid them in evaluating site usage so that they can identify problem areas and alter them accordingly. At present, the most popular tools utilised in this evaluation make use of a technique called log file analysis, a process by which server log files are parsed to extract information about visitors to a site. However, when visualising this information, current tools have either neglected site structure or else failed to utilise properties unique to web sites. We address both these issues by developing a visualisation of server log data that incorporates site structure and caters specifically for web sites by exploiting their unique characteristics.