[15598] in Public-Access_Computer_Systems_Forum
Web page co-use information/analysis?
daemon@ATHENA.MIT.EDU (Theodore Morris)
Mon Nov 22 20:12:18 2004
Date: Mon, 22 Nov 2004 13:59:34 -0500
From: Theodore Morris <tamorris@KENT.EDU>
To: PACS-L@LISTSERV.UH.EDU
Reply-to: Public-Access Computer Systems Forum <PACS-L@LISTSERV.UH.EDU>
Message-id: <5.1.1.6.0.20041122134625.03f03a40@pop.kent.edu>
MIME-version: 1.0
Content-type: text/plain; charset=us-ascii; format=flowed
I want to do a webpage co-occurrence analysis for a colleague's website.
I'm basically coming from a co-citation/co-occurrence background, so that's
framing how I see the problem.
From the page-co-occurrence analysis, we would hope to be able to learn
more about what combinations-of-pages are being used most often. We're
interested in finding software (freeware preferred, of course, but $$ware
also considered!) that will facilitate going through web site transaction
logs to find "what pages were used with what other pages?"
Naturally, we expect the major menu pages, etc., to be used most
frequently, and in fact my colleague's current weblog analysis package
already gives a report of page requests by frequency for a given time
period. What we're looking for now is to follow an IP-address'
page-requests through a presumed-contiguous thread of requests representing
a single search session, to observe usage patterns.
I'm not seeing much of anything in the journal/web literature on this--but
as I say, I may be looking in the wrong places/using the wrong buzzwords.
Has anyone attempted this before, or can anyone direct me to some likely
useful resources to kickstart my search? Thanks!
Ted
Theodore Allan (Ted) Morris, PhD | tamorris@kent.edu
Asst. Prof., Columbus Pgm., Kent State U. | wb8vnv@hotmail.com
School of Library & Information Science | "Hey, TED!!"
124 Mount Hall - 1050 Carmack Rd. | Voice Fax
Columbus, OH 43210 | 614-292-7746, -3618
Failure isn't the falling down, it's the not getting up again.
