Re: [analog-help] Report results comparison (Analog vs Web Trends)

2002-04-03 Thread Jeremy Wadsack


Berigan, Matthew ([EMAIL PROTECTED]):

 Hi!

 I'm finally past bad log data and I think that I'm getting inconsistencies
 between WebTrends and Analog that can't be explained by INCLUDES or
 EXCLUDES.

 First, here's the summary results:

 ANALOG
 Program started at Wed-03-Apr-2002 09:38. 
 Analysed requests from Sun-24-Mar-2002 00:04 to Sat-30-Mar-2002 23:58 (7.00
 days). 
 Successful requests: 76,490 A
 Average successful requests per day: 10,933 D
 Successful requests for pages: 25,765 B
 Average successful requests for pages per day: 3,682 C
 Failed requests: 760 
 Redirected requests: 680 
 Distinct files requested: 2,047 
 Distinct hosts served: 5,009 E
 Corrupt logfile lines: 4 
 Data transferred: 792.723 megabytes 
 Average data transferred per day: 113.313 megabytes 

 WEBTRENDS
 Date  Time This Report was Generated Wednesday April 03, 2002 - 09:17:45 
 Timeframe 03/24/02 00:04:41 - 03/30/02 23:58:54 
 Number of Hits for Home Page N/A 
 Number of Successful Hits for Entire Site 77138 A
 Number of Page Views (Impressions) 25816 B
 Number of User Sessions 8488 E
 User Sessions from United States 0% 
 International User Sessions 0% 
 User Sessions of Unknown Origin 100% 
 Average Number of Hits per Day 11019 D
 Average Number of Page Views Per Day 3688 C
 Average Number of User Sessions per Day 1212 
 Average User Session Length 00:08:36 

 Most everything (A-D)appears to be close enough that I don't even consider
 it worthy of question.

Well, to answer anyway, WT counts some redirects as successful
requests that Analog does not.


 Item E, however, bugs me. Can anybody guess why
 Distinct Hosts [Analog] and User Sessions [WT] should be so different?

Because they are different quantities.


 I think this is supposed to be the same data and my assumption is
 that it is the number of distinct IP addresses listed in the log (I
 suppose I could count those (ck!!)).

Sessions is something that WT estimates (and by most accounts very
poorly.) It does not represent the number of distinct IP addresses in
the logs. It is meant to represent the number of distinct user
sessions. Read http://www.analog.cx/docs/webworks.html as to why
that's not easily counted and why Analog does not do it.



-- 

Jeremy Wadsack
Wadsack-Allen Digital Group

+
|  This is the analog-help mailing list. To unsubscribe from this
|  mailing list, go to
|http://lists.isite.net/listgate/analog-help/unsubscribe.html
|
|  List archives are available at
|http://www.mail-archive.com/analog-help@lists.isite.net/
|http://lists.isite.net/listgate/analog-help/archives/
|http://www.tallylist.com/archives/index.cfm/mlist.7
+



Re: [analog-help] Report results comparison (Analog vs Web Trends)

2002-04-03 Thread Stephen Turner

On Wed, 3 Apr 2002, Duke Hillard wrote:

 In a Unix shell environment (including Cygwin), it's not
 difficult to count the number of unique hosts from a log.
 Something similar to the following should work fine.
 
 cut -d  -f1 /dir/subdir/server.log  /dir2/somefile.txt
 sort -u -o /dir2/anotherfile.txt /dir2/somefile.txt
 grep -c . /dir2/anotherfile.txt
 

But note that this will count hosts with only failed requests. The number
analog reports is the number with successful requests.

-- 
Stephen Turner, Cambridge, UKhttp://homepage.ntlworld.com/adelie/stephen/
This is Henman's 8th Wimbledon, and he's only lost 7 matches. BBC, 2/Jul/01

+
|  This is the analog-help mailing list. To unsubscribe from this
|  mailing list, go to
|http://lists.isite.net/listgate/analog-help/unsubscribe.html
|
|  List archives are available at
|http://www.mail-archive.com/analog-help@lists.isite.net/
|http://lists.isite.net/listgate/analog-help/archives/
|http://www.tallylist.com/archives/index.cfm/mlist.7
+