Virtual visitors to the SLV website

In the words of its own website, the State Library of Victoria is 'the major reference and research library in Victoria, responsible for collecting and preserving Victoria's documentary heritage and making it available through a range of services and programs'. While the physical space of the library has always been an important aspect of the public library (Council on Library and Information Resources 2006), in recent years, like many libraries, the State Library of Victoria also has a virtual presence. The SLV first had a website in the mid 1990s. and the current website, referred to in this paper as slv, has been online since 2004. While in 2006-07, there were just over one million visits to the buildings of the State Library of Victoria, in the same year, there were more than triple that number of visits to its website and more than 22 million visits to websites it supports. Although the library regularly conducts surveys of the physical visitors, to find out who is visiting the library, the reasons for their visits and their satisfaction with library services, little is known about the virtual visitors to the library. There is an enormous amount of SLV web log data that can be used to analyse the activities of online visitors; some of it is free, some of it available for a fee. Although the reporting of this data is usually tailored for market research type purposes, it has largely untapped potential for finding out more about the visitors to library websites. Recognising this, Jones et al. (2004) have outlined a program for analysing web analytics of a digital library and Fang (2007) uses Google Analytics to show movement within the site of a particular digital library. The vast quantity of data can be overwhelming and so there have been a number of attempts to develop sophisticated automated processes for analysing web log data (for example, Norguet et al. 2006). This paper differs from most studies in two ways. In order to find out more about how SLV is meeting users' information needs, this paper attempts a manual classification of the long tail of upstream and downstream websites. While there has been a deal of analysis of the structure of search queries, and the amendments made (for example Hargittai 2004; Spink and Jansen 2004) there has been little analysis of their semantic content, except in cases where this is limited to very specific subject domains. This paper, however, includes an analysis of the long tail of search queries in order to understand more about the type of content people are accessing on the site.