How do I achieve point in time analysis in a datawarehouse

  • Hello all

    I have created a prototype datawarehouse for my company using Analysis services 2005. The data source view is in a star schema and the cube has the following facts and dimensions.

    My fact table is called Logins it stores the userId of the user that logged in and login date/time. I have a dimension called Member which contains attributes about my member and a server time dimension.

    So a query that we could run is a break down of how many Active members logged into the system in 2007 broken down by month. So the counts would be displayed for each month for the year 2007. The problem with this is that the members active attribute might have changed after 007 therefore the counts would be different. So in order to get the counts that would have been displayed if I ran this query at the end of 2007 I would require the member dimension as it stood then.

    Is there anyway to achieve this in Analysis Services 2005?

  • Hi Peter,

    I suggest you have a 'Member status' dimension. you can always get the answer for 'number of people logged in to the system in August 2007' but selecting both 'Active' and 'Inactive' members. At any point of time, you can also find 'how many members were present in August 2007' but not today.

    let me know if this would work for you..

    Warm Regards,
    Neel aka Vijay.

  • Hello Neel

    Thank you for your response your solution will work. The member dimension has to many attributes as it stands and should potentially store some of that information in the fact table every time a login occurs. This way when attributes change for a member it will be reflected in any query because the information would be stored at the point in time of the login.

    Thank you once again

    Peter

Viewing 3 posts - 1 through 2 (of 2 total)

You must be logged in to reply to this topic. Login to reply