When you decide which KPIs you wish to measure and according to which criteria to slice & dice them, you must be aware of the following table types.
When you start to import your data into the Elasticube, you need to differentiate between the table types. This is the basis for a successful data model and for understanding of the data.
Fact – Usually it is a transactional table which contains the fields we would want to perform our calculations upon (measures). Additions/Updates are performed at a high rate.
- Examples: Tickets, Sales, Inventory levels, Calls, etc.
Dimension (Dim) – Descriptive tables which contain the attributes which the measures can be sliced & diced upon. Additions/Updates are not frequent.
- Examples: Products, Customers, Date, Countries, etc.
Note – Sometimes we will use the Dimension tables as a source for a measure, e.g. count(CountryName).
Connection/Mapping Table – Consists records of association between two tables. Which Id from table A is associated to an ID from table B (many to many by design). Measures in the front end may need to be tweaked by using more complex calculations.
- Examples: Which student is in which course, which project contains which tasks, etc.