Cubes without a Data Warehouse? How?

Perceived wisdom says that in order to deliver OLAP cubes you will need a Data Warehouse. Indeed in some circumstances this remains the case, and in the "Personal BI, a positioning statement" page we explore the specific circumstances where either is true. All that said, the fact remains that it is not entirely necessary to have a full data warehouse in order to enjoy the benefits of OLAP.

This is the essence of Personal BI. The ability for the individual to benefit from being able to ask questions of the data they have available, turning that data into information quickly and easily, thus helping to make business decisions.

One of the products you will see available on this site is CUBE-it Zero, a free OLAP cube builder that builds OLAP cubes without the need for a Data Warehouse implementation.

By simply looking at the data you have available to you, and categorising the elements (fields) into things you want to measure (count, calculate etc) and things you want to analyse by you have all you need to achieve building an OLAP cube.

Let's take a simple example.

Dave is a sales manager, and every week he get's a report of sales it shows Total sales summarised by branch and region. He also get's a report which shows the top ten customer sales and which salesperson is responsible for those customers. He is also copied in on a monthly report which shows product performance by product category. These are all really interesting reports, but one of the many questions he can't answer is what products are the top ten customer buying and also what products are the sales team successfully selling, and equally unsuccessfully.

The answers to these key questions, help to determine activity for the coming weeks/months.

Personal BI, can deliver the answers to these questions, and given that the reports are already being produced and the data is available, derriving a cube is simple enough, the current reports provide

Customer, Salesperson, Branch, Region, Branch City, Product, Product Categrories, Sales Value, Product Quantity Sales.

To generate a cube using CUBE-it Zero from this we simply identify what we want to analyse by and what figures we want to analyse

All of the columns we want to analyse by become "Dimensions", and the figures we want to analyse become "Measures". This is all we need to define a cube in CUBE-it Zero, and hence we derrive a cube without a Data Warehouse.

The built cubes are used to feed dynamic reports produced in Dave's reporting tool of choice*, that not only answer the questions Dave needs to ask, but also provide interactivity using drill down etc.

* CUBE-it Zero Professional can use SQL Server Analysis Services to produce OLAP Cubes, as such any reporting tool such as Business Objects, COGNOS, MS Excel, Crystal Reports etc etc, can be used to produce reports. Other editions of CUBE-it Zero (Foundation and Personal) do not use Analysis Services and therefore support the production of local cube files.