7 thoughts on “Tabular Models, Multiple Data Sources and Unknown Members

  1. Paul

    i really appreciate your article, but coming from a “power” Excel user background, SSAS tabular was the best thing that happen to me professionally, from doing rigid error prone massive spreadsheet to the stage where I built my “personal” BI semantic layer that delivered reports used and appreciated by my clients.

    I understand the need for a DW, data governance, doing ETL at a first stage, i really do, but we have to work with what we have which means just software installed in our laptop, no IT, no corporate BI, no training nothing,

    basically the DW and the semantic layer are built in the same data model using PowerBI desktop, the data sources are a combination of operational database and flat files, the reports are authored in Excel using a hack to connect to Power BI desktop as the cloud is not an option.

    yes it is slow sometimes, the model need to be refreshed manually, but it is much better than using Only traditional Excel or God forbidden MS Access, and it is flexible, if someone want a new dimension fine, just fire a new PQ connection find a common key and it is done.

    so for the real benefit of PowerBI stack is to replace spreadsheet hopefully not corporate BI solution (if they exist).

    happy new year

    1. Thank you for your thoughts on this. You have reaffirmed what so many others are experiencing. Self-service BI and analytics, in many places, are becoming the normal mode from running business. I think more self-made BI practitioners like yourself recognize the value of “doing it right” at a corporate or organizational level vs the reality of “just getting it done” with tools that are available, like Power BI Desktop.

  2. Great article. With the release of SSRS 2016 the MS BI stack gives us a set of tools to handle all disparate data sources within organisations and satisfy both cloud and on premises environments.
    I agree with Chris its not either or….as PowerBI brings us an even ‘quicker’ platform.

    MS BI stack link : http://blogs.technet.com/b/dataplatforminsider/archive/2015/10/29/microsoft-business-intelligence-our-reporting-roadmap.aspx

    1. Ah, but with great power comes great responsibility. I agree with you that the reporting platform is evolving forward in some promising ways. With so many choices, how do you guide less-experienced practitioners to choose the right tools & architecture?

  3. Interesting points. I don’t see Power BI etc as a replacement for a full DW architecture, more a complementary offering. The full DW architecture provides most control, options, etc but is also the slowest to deliver. In most organisations there is a limited ETL/DW resource pool and a large backlog. This means lower priority requitements, even if somewhat simpler, never get met. Power BI, etc is often a suitable tool for these. IT can manage the staging/ODS but are not the bottleneck thereafter. In lifecycle terms, as the understanding and value of a data source/set rises, a decision can later be made to bring it into the DW. Power BI etc can still be used against a DW as well. So I don’t see this as either or. Ultimately a DW is the most capable solution, but it doesn’t meet the needs of every set of requirements at every point in time. The new tools provide new options that need careful decisions about where and when they are most appropriate.

    1. Thanks, Chris for your insights. A point that I wanted to emphasize is that since the newer modelling tools provide less capability for Unknown member handling, at what point do you push-back on an existing model design and insist on a multi-tier redesigning?

Leave a Reply

%d bloggers like this: