Nicky and Jefferson attended the Wherescape seminar yesterday morning in the 17th floor penthouse at the New Zealand Embassy just off the Pall Mall. It was a bright and misty autumn morning which added to the atmospheric view across central London.
There were about 40 people there and the main part of the morning was headlined as a demonstration from one of Wherescape’s consultants of how to create a data mart in an hour using Wherescape’s product “RED”. In fact including explanations to the audience it took about 45 minutes. I have summarised what was covered in the demonstration below.
There was also a presentation from Lawrence Corr, a data warehouse designer Jefferson has worked with before and who has just published a book on agile data warehouse development, due out in the next two weeks. We look forward to reading it Lawrence!
The final part was a presentation from Shawn Lewis, BI manager at Vodafone UK, talking about how Vodafone are using RED in several countries as
a prototyping tool alongside the more technical track of Ab Initio and Teradata.
Overall we thought it was a really good event, well done to Rob and Emma for organising – and for anyone who was there, did you have any idea that Rob was so into Star Wars?
We’re running a short series of blogs on data warehouse development and agile in particular, the introduction to the series is here: http://bit.ly/nJ6xPp and the follow-on focuses on what you need to have in place as an organisation to really benefit from it.
The main things covered in the RED demonstration were:
How the product itself is organised for use by data warehouse designers. The main groupings are Connections, LoadTables, Dimensions, Dimension View, Stage Tables, Fact Tables, Aggregates, OLAP Cubes, OLAP Dimensions, Export, Procedures, Host Scripts and Indexes.
Connecting to a data source. In this case he connected to a Microsoft standard data set in SQL Server, but the product connects to any ODBC or OLE-DB data source and itself runs on SQL Server, Oracle, Teradata or DB2. Support for database appliances is under development.
Using the Load Wizard to extract data. He also later showed how that could be scheduled by time or by triggers.
Creating dimensions. A wizard allows you to select Kimball type 1, 2 or 3 and the appropriate code is generated in the background. Loading can be cursor or set based. A business key is selected when creating a dimension and a surrogate key is automatically created in the background. He also showed that indexes are automatically createdon the surrogate key and the business key.
Setting up a fact table and then drag and dropping using a visual interface to create a star schema connecting the fact table with dimensions.
Automatic lineage availability and self-documentation, both technical and business.
Creating OLAP cubes and setting up hierarchies on dimensions for the cube, via a wizard.
Changing a dimension at a later date, from type 1 to type 2.