Free
Web Mining
tool

VISITaTOR

Clustering and visual presentation of visitor groups
based on access patterns

forum on reatika.ru

Open the static result page in Internet Explorer and read the comments and explanations below.

This is result of clustering for the site forum.reatika.ru. The language of the site is Russian.
The navigation of the site occurs partially per parameters in the searchstring, and partially per static links.
Because this version of VISITaTOR ignores the parameter string, the clustering describes the types of acts (login, profile etc.), not the groups of interests. Only a small percentage of clusters - the ones where the navigation occurs via simple links - reflects the areas of interests.

From bottom to top:

  • programming, PHP - the main URLs of this group - forum-3.html and about23.html - deal with programming and PHP;
  • profile - because of the aforementioned absence of the parameter string we can only suppose that these visitors come to the forum to alter their own profile or view the profiles of others
  • business - the main URLs of this group - forum-1.html, about9.html and about36.html - deal with Internet business.
  • rss - the users who access the rss feeds via web browser. On the Browser table we can see that the group is based on regular accesses of one person who uses Opera Browser
  • site creation, beginner - topics about site creation for beginners
  • posting - Yandex - the users come from the Russian search engine Yandex (we can see it in the Referrer table) to read a text found through their queries. They do not stay on the forum to read more.
    In the next group we will find the subgroup "viewforum,viewtopic,posting - Yandex" - probably these are the users from Yandex who got interested in the forum.
  • the main cluster "viewforum,viewtopic,index" - because of the absence of the parameter string we cannot recognize the interest areas of this group.
    In the Referrer table we can see that the most accesses occur without referrer, that is to say, from bookmarks or prompt in the URL field of the browser. This could mean that these are not random but regular (propably registered) users.

    So we can suppose that at least half of visitors on this site are regular.

The distribution of groups with known topics (7% site creation - 2% business - 1% programming) indicates that it is the same among registered users, but the size of these groups is too small to accept it.
Feel free to contact me if you are interested in developing this tool into a commercial product.   © Natalia Bazhenova 2005 
 
Visit my sites:
Immobilien ohne Makler | Immobilier sans courtier | Real estate with no agents | AuctionX