Ralph Kimball’s Bottom-Up Business Intelligence Design, a renowned author and expert in the field of data warehousing. He is a proponent of a bottom-up approach to data warehouse BI design. This is the “Bottom-Up” method. In this approach, IT creates data shops or data marts to provide dedicated reporting and analysis capability. For specific business processes, and are easier to use than complex data warehouses.
Presentation of Ralph Kimball’s bottom-up business intelligence modelling methodology
In Ralph Kimball’s methodology, the bottom-up process is the result of an initial undertaking to analyse the relevant business processes to be modelled.
Specialised data shops
Data shops contain the dimensions that are the focus of analysis and the facts or measures. Facts can contain either atomic data, i.e. at a fine level, or aggregated data. It often models a very specific area of activity such as sales or production. People involved integrate these data shops with a business intelligence solution to create a complete datawarehouse.
This integration between different data warehouses is implemented by what Kimball calls a data warehouse bus architecture. It is enabled by a collection of conforming dimensions, which are dimensions that are shared (in a specific way) between two or more data repositories, allowing cross-analysis across multiple business areas or processes.
Raplh Kimball’s Drill Across
Cross-data integration in the data warehouse is based on conforming dimensions that represent entry points between data marts. Teams do the actual integration of two or more data marts through a “Drill Across” process. It is a lateral drilling that brings together different business data but at the same level of granularity and using the same dimensions. The dimensions often used because they are cross-company are, for example, Time, Customer or Product.
Maintaining accurate management of the data warehouse architecture is essential for data integrity. The most important management task is to ensure that the dimensions between the data marts are compatible and therefore updated in parallel. Some analysts believe that it is an advantage of the Kimball method that the data warehouse ends up being segmented. Indeed, into a number of logical and coherent data marts, rather than one large, centralised and often complex model.
An iteration-friendly connected silo view
Business analyse the company’s data can be as soon as the first data shops is available. The method allows an exploratory and iterative approach to building the warehouse. For example, teams develops the data loading effort for sales, with a specific data warehouse.
The rest of the BI project can continue for Production. And a joint analysis can highlight the correlation between production capacity and daily sales. This Ralph Kimball view of the BI world is conducive to iterations. And also allows the production of BI projects domain by domain.