1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

I need web logs that contain sessions data, you NEED to know user information

Discussion in 'BlackHat Lounge' started by IamNomad, Jun 5, 2011.

  1. IamNomad

    IamNomad Junior Member

    Joined:
    Mar 27, 2009
    Messages:
    135
    Likes Received:
    22
    Location:
    ┌∩┐(◕_◕)┌∩┐
    Heres the deal, for a Grad class i wrote a framework to analyse apache based web logs and apply heuristics to the data to get meaning full results back. I've done as much testing as i can from data that i own. I'm looking for someone who owns an ecommerce site or a large site that has a particular advertising campaign. (perhaps a CPA offer 3 levels deep).

    So what do you get for your generous donation? Ill give you a summary of the rules generated by my framework. FOR FREE. This type of thing is valuable to tuning a campaign , restructuring a website, product placement etc.

    I only have the time to pre-process and analyze one good set of logs this weekend. so this is first come first serve.

    what makes up a good log:
    -website that get SOME traffic - empty logs do me no good
    -logs that have session IDs written to them in apache - they dont necessarly have to contain this, but it make the results hella more accurate.
    -at least 30 days of traffic for me the analyze - the size of logs can be in the gigs depending on your traffic.

    The logs will need to be put on an ftp , mediashare, or http server somewhere where i can download them.

    PM me w/ intrest.

    note: admins i thought the lobby would be the best place as im not selling a service and not asking to JV. if i posted in the wrong spot please move somewhere appropriate.
     
    Last edited: Jun 5, 2011
  2. IamNomad

    IamNomad Junior Member

    Joined:
    Mar 27, 2009
    Messages:
    135
    Likes Received:
    22
    Location:
    ┌∩┐(◕_◕)┌∩┐
    I got a PM what is actually involved here. so ill bump my own thread to answer.
    the apache logs is what i would looking at : /var/log/httpd or logs>raw log file if your using cpanel

    I would then run my preprocessing script to format the data properly and my data mining suite when then chew on it and run quantitative models against and spit out whatever rules it can derive from the info

    For a example, results like :
    50% users who actually bought something took path a>c>d>checkout
    Users who did c>b bounced/left the page
    All users who went to 78% of users who went to E bought something

    we might be able find something like:
    Users therefore we should look at B and see why people bounced. How did they get there id they didn't start at A?
    Whats special about E? did we do some marketed or SEO campaign specifically for E. Did some referrer like us and we got a crazy amount of sales ?