This is an area that I have not yet started on.

The idea of log 'analysis' will lie in three areas:

1. Scanning the raw log files and/or scanning the delimited files (prepared for
database insertion) for items of interest.

2. The interface between the presentation server (httpd) and the RDBMS. All of
the queries built by this interface constitute the analysis of the log data in
the database.

3. Queries submitted to the RDBMS (probably as 'cron' jobs) to retrieve 
data for statistical and/or other analysis.