Archive

Archive for the ‘Data Modeling’ Category

Tableau 10 Cross Join – An alternative way to achieve CrossJoin / Generate in MDX / DAX

November 12, 2016 Leave a comment

Just watched the Tableau Conference 16 session “Cross Database Joins: The unexpected solution to many tough analytics problems”. It is an awesome session, solved some problems which can not be solved by Tableau (before V10) alone.

Essentially, it is trying to solve the famous “event in progress” or “events with duration” questions. This kind of questions can be solved in SQL Server Analysis Services Multidimensional mode, Tabular mode, Power Pivot or Power BI since the beginning. Having said that, this also means the modelling capability of Tableau is one step closer to the OLAP (such as Power BI or Power Pivot for Excel).

A side note, now, within Tableau’s world, the near impossible accumulative distinct count should be able to achieve by this cross database join approach.

“future” approaches to BI/DW implementations

March 19, 2010 Leave a comment

Today I read a very interesting article about "future" approaches to BI/DW implementations which can be found at here.

These days I am working on an enterprise data warehouse for a major bank in Australia. My team has a lot of debate around using Generic Pattern / UDM (Len Silverston’s Universal Data Model) or Data Vault. One major advantage of UDM here is the larger deployment in industry. There are more case study and existing model template made available for quick deployment.

Currently, we adopt Agile methodology for the project. We use Test Driven Development (TDD). As the database platform is IBM DB2 9.7 and Oracle 11g, we use  Concordion automated unit testing framework. And recently I found on CodePlex, there is a free unit testing framework for Microsoft BI, which is called BI Quality, which is based on NUnit.

Categories: Data Modeling