hello to all,

please let me report my experience with phpopentracker. i tested it with a
lokal copy of fluter. i made a template called tracking.tpl which contains
the empty-gif-php-script of phpopentracker. i included this template in the
main article.tpl. so for every response for an article the tracking engine
is working. the standard tracking entries are stored in the tables
"accesslog" and "documents". "accesslog" contains only integers, "documents"
contains the url of the document. the connection between "accesslog" and
"documents" is the data_id of the requested document.
the data_id is an integer encryption (crc32 polynomial) of the url-string.
so the engine doesn't need to check whether the document is already tracked,
it just makes an entry with this encrypted data_id and it gets only unique
entries. i think, this is a very performance way to handle the queries.
there is another database table called "add_data". beside the standard data
like referers, user agents, visitors, hostnames and operating systems, you
can store any kind of data to this table. so i use this container to track
especially campsite data. these are: IdLanguage, NrIssue, NrSection,
NrArticle. please let me know, if there is one id missing. i catched the
data in a separate database table using plugin-interface of the
logging-engine. i think, this is an important information for a specialized
campsite view like top article or main language.

i didn't test any other tracking tool, but i really get along with
phpopentraker very well. it is well documented and clearly coded. so i would
prefer to implemate phpopentracker in campsite.
there is a really good api for rendering the tracking data. you can use
JPGraph, a php-wrapper for graphic visualisation. here are some points of
the api calls index:
-num returning visitors
-num unique visitors
-page impressions
-returning visitors
-top paths
-average time between visits
-inividual clickpaths
i think, these points are covering the main requirements of the document
below.

some points to campsite-architecture:
i installed phpopentracker as a pear-module. so the files are stored in the
pear-directory. i'm not sure whether i can use this for a global campsite
installation.
all other data is stored in the priv directory in a subfolder called
"tracker". is this ok?
i need a global php-file for including all information concerning campsite
like database name, user, password and so on.

to do:
-we will have a betatest with the tracking-template on the
fluter-livesystem. so we can check the performance of this tracking tool. we
will test the tracker with realtime database inserts and offline inserts
(the textfile solution)
-interface for the api and connecting it with the campsite-admin interface

best, flo

> Traffic measurement software: Most wanted capabilities:
>
> This document lists both the urgently needed capabilities we need from a
> trafic
> measurement application as well as the more advanced options which, though
> not
> strictly necessary would be very nice to have.
> The basics
> -Number of daily/weekly/monthly visitors
> -Exact number of daily/weekly/monthly page views
> -More detailed breakdown of readers by country
>
> Current application shows data that does not seem to correspond with banner
> serving software and we suspect that data is incorrect and somewhat inflated.
> -An exact breakdown of most popular articles by week/month
> -An exact breakdown of most popular sections
> -Number of page views per reader
> -Number of unique visitors as opposed to only measuring total visits
>
> TOL's current software cannot filter data to show which articles and sections
> are the most widely read. This data would help TOL get a sense of what sells
> and
> what our readers prefer, and could TOL help structure its site better while
> focusing its attention on sections that are neglected.
> More advanced
> -How long is the average user session?
> -How often does the reader come back to the site?
> -How many visits from subscribers as opposed to free readers?
>
> This would help us understand how many subscribers visit the site, and how
> often
> and what their proportion of overall readership is.
> -How many readers click on printer friendly as opposed to clicking their way
> through the article?
> -How many readers click through to the last page?
> -Most frequent entry pages
> -Most frequent exit pages
>
> This data can greatly help us understand what pages users enter TOL through,
> which would help us focus our efforts on that page(s) trying to make them
> more
> sticky and easy to use as a springboard deeper into the site.
> -The ability to cluster countries and measure traffic for any country or
> group
> of countries
>

------------------------------------------
Posted to Phorum via PhorumMail