ktmatu
Home | Site Map | Site Search
Home > Software > Dotlog Analysis Tools >
 

Dotlog Analysis Tools Examples

This page gives some examples of how to analyze log files generated with the cookie logging feature enabled Blue dot script. Examples require that you have Dotlog Analysis Tools installed on your computer.

It is assumed that your log file is called accesslog. Commands that you are supposed to type in on the command prompt are marked like this. Computer's responses look like this.

Commands like sort, wc, uniq, awk, xargs, grep and less are usually available on all Unix systems. GNU utilities for Win32 provides these commands for Windows users.

How many unique browser sessions?

% ./sidextract.pl accesslog | sort -u | wc -l
    4682

The number of unique browser sessions is approximately the number of unique visitors.

How many unique browser sessions in June 2003?

% grep /Jun/2003 accesslog | ./sidextract.pl | sort -u | wc -l
    3516

On average how many page views per session?

% ./sidextract.pl accesslog | sort -n | uniq -c | sort -n | awk '{s += $1} END {print s / NR}'
1.78941

This statistics can be used to detect click fraud.

How many sessions generated by repeat visitors?

% grep 03/Jul/2003 accesslog | ./repeatvisits.pl | wc -l
       3

In this case the repeat visitors are those who have visited the site sometime before July 3, 2003.

Log entries generated by repeat visitors

% grep 03/Jul/2003 accesslog | ./repeatvisits.pl | ./pidextract.pl | xargs -I '{}' grep '{}' accesslog | less
...

The same in two steps:

% grep 03/Jul/2003 accesslog | ./repeatvisits.pl | ./pidextract.pl > pids.txt

% grep -F -f pids.txt accesslog | less
...

N.B. Solaris grep does not support -f. Use /usr/xpg4/bin/grep or GNU grep instead.

What are the entry pages of those who visited thanks-for-buying.html page?

% grep thanks-for-buying.html accesslog | ./pidextract.pl | xargs -I '{}' grep '{}' accesslog | ./firstvisit.pl | less
...

Entry page is the page by which visitors enter your site.

These log entries describe the source of traffic. If you are using pay-per-click search engines, this is very valuable information when you are calculating your ROI or selecting which search engines and keywords to use.

Out of this data, Relax can create a report where referring keywords, search engines and URLs are listed in detail.

What are the log entries generated by those who have not visited thanks-for-buying.html page?

% grep -v thanks-for-buying.html accesslog | ./pidextract.pl | sort -u > nobuy.txt

% grep thanks-for-buying.html accesslog | ./pidextract.pl | sort -u > buy.txt

% comm -12 buy.txt nobuy.txt > buyandnobuy.txt

% grep -v -F -f buyandnobuy.txt accesslog | less

Out of this data, Relax can create a report where referring keywords, search engines and URLs are listed in detail. This makes it easier to understand why your visitors have not taken the desired action.

N.B. Solaris grep does not support -f. Use /usr/xpg4/bin/grep or GNU grep instead.

What are the exit pages on a particular day?

% grep 03/Jul/2003 accesslog | ./pidextract.pl > pids.txt

% tail -r accesslog | grep -f pids.txt | ./firstvisit.pl | less
...

Exit page is the last page people see before leaving your site.

N.B. Solaris grep does not support -f. Use /usr/xpg4/bin/grep or GNU grep instead. GNU tail does not support -r, use tac or perl -we 'print reverse <>' instead.

On average how many pages are viewed by those who visited landing-page.html?

% grep landing-page.html accesslog | ./pidextract.pl | sort -n | uniq -c | sort -n | awk '{s += $1} END {print s / NR}'
1.16493

Landing page is the page on a website where the visitor arrives. This statistics can be used, for example, to detect click fraud.

 
Home | Software | Information | Etsin | Chinese | Christmas Calendars | Site Info