|
Just like Facts+ Derived Facts matrix table, we also need to now start working out the dimension + attributes table. This is a key building block for dimensional modeling as you are creating 'Foundation' dimension. For example a 'Customer Dimension' will be a reusable dimension, which can be used in Sales Revenue analysis, Customer Service management, Customer Value and Profitability analysis and so on.
To create a dimension, one has to look at all aspects OR possible uses it can be put to. This will be helped by the 'Data Marts+ Dimensions + facts matrix' where a dimension's application to various kind of analysis will be mapped.
It may be easy to ensure that all possible dimensions have been listed out. How do you make sure that all possible attributes to a dimension have also been listed out? Following are the ways, which may help you: - Pedantic method of picking up the fields to which the primary key of the dimension belongs to. For example consider all fields from 'customer_master' for customer dimension.
- Review the data marts and by looking at the analysis, add additional attributes. For example for portfolio management, one may like to place the 'risk category' as the additional attribute.
- Check industry best practices.
Another example can be to place the 'customer segment' code as an attribute.
As you refer to business hierarchies in OLAP, you can understand that the various levels in the hierarchy are marked as attributes. For example, the location dimension will have the 'Office' as the most granular element (or instance). The hierarchy attributes of the location dimension will be office PIN code, Office city, Office district, office state, office country. The attributes, which are used for analysis are called 'classification' attributes. the hierarchy related attributes are one kind of 'classification attributes'.
PLEASE REFER Execution-MiHPractice Tool Dimensions & Attribute listing |
