<Home

ask us!

email | chat | phone

UW WorldCat
[advanced search]

UW Catalog
[advanced search]

Site Search

Collection Assessment from Every Angle

WebBridge Log Analysis

By analyzing our WebBridge link resolver logs, we are able to identify patterns of user behavior that is not available from vendor supplied usage statistics.

The log files generated by the WebBridge link resover provides data on linking origin, linked resource, a date and time stamp and whether the link was clicked. This information is enough to identify some usage patterns of both origins and resources that is difficult to get from any other source.

The raw log file looks something like this:

STATISTICS from 06-01-08 to 06-30-08
RESOURCE ORIGIN PLACEMENT DATE TIME CLICKED
Academic search complete ISI includes Web of Science (V 1.0) Resolution server 06012008 000506 1
Academic search complete ISI includes Web of Science (V 1.0) Resolution server 06012008 000506 0
Academic search complete OCLC Worldcat Local Resolution server 06012008 000752 1
Academic search complete OCLC Worldcat Local Resolution server 06012008 000752 0
Science Direct ISI includes Web of Science (V 1.0) Resolution server 06012008 151916 1
Science Direct ISI includes Web of Science (V 1.0) Resolution server 06012008 151916 0
Science Direct ProQuest (All) Resolution server 06012008 152006 1
Science Direct ProQuest (All) Resolution server 06012008 152006 0
Science Direct Gale Resolution server 06012008 152038 1
Science Direct Gale Resolution server 06012008 152038 0
Science Direct Google Resolution server 06012008 152131 1
Science Direct Google Resolution server 06012008 152131 0
Science Direct Google Resolution server 06012008 152330 1
Science Direct Google Resolution server 06012008 152330 0
Science Direct Ovid Resolution server 06012008 152649 1
Science Direct Ovid Resolution server 06012008 152649 0
Science Direct PubMed Resolution server 06012008 152727 1
Science Direct PubMed Resolution server 06012008 152727 0

A Perl script is used to parse the log file, reformat the date and time stamp and upload the data into a MySQL database running on a web server. For some limited analysis, a desktop database such as Microsoft Access can be used, but for anything more than a couple of months worth of data, a more robust database is needed. The current database in MySQL contains about two years of data and is up to a little over 4.5 million rows.

With the WebBridge data in the database, we can run a series of SQL queries to extract a variety of information. For the quarterly summary report prepared for staff and administrators the tables and data include (see examples below):

• List of top ten origin sources for each month and for the quarter;
• Table and chart showing the number of origin links coming from WorldCat Local;
• Table showing the top resources for origin links coming from WorldCat Local;
• Table and chart showing the number of oriting links coming from GoogleScholar;
• Table showing the top resources for origin links coming from Google Scholar;
• Table showing the top resources links for all incoming links.

Additional custom queries can be run to find virtually any combination of origin and resource, for example, top origin sources for JSTOR, or top resources for Web of Science. The WebBridge logs do not store the information needed to do more discrete analysis such as link activity to individual journals or requests to journals that we do not have available full text. Some of these features are available from other link resolvers such as SFX.

Examples

2007 July 4653 2007 August 5193 2007 September 5874 2007 October 15566 2007 November 17819 2007 December 9762 2008 January 12567 2008 February 19214 2008 March 18507 2008 April 19768 2008 May 25019 2008 June 11941

2006 2007 2008 January 1583 2576 February 2473 4344 March 2768 2856 April 3457 3592 May 5233 June 2363 3162 July 2069 August 2183 September   1553 October 238 2914 November 805 2895 December 1040 1535

Last modified: Monday August 04, 2008