What is a dimension hierarchy?
A dimension hierarchy is a data construct used to group data into bins based on value ranges. A dimension hierarchy can contain multiple levels, each of which has its own set of bins. The bins in each lower level must roll up neatly into bins in higher levels.
For example, the Age dimension hierarchy could have two levels, Lowest level and Rollups. Customers are grouped in bins at each level:
Lowest level:
(21-25), (26-30), (31-35), (36-45), (45-59), (60+)
Rollups:
Young (21-35),
Middle
(36-59),
Older
(60+)
You cannot split a lower level bin (for example, bin 26-30 above) and divide individuals aged 26-27 into "young" and 28-30 into "middle", when rolling up to a higher level. Any single bin in a lower level must fall completely within a higher level bin. If you wanted to actually define "young" as those aged 21-27, you would need to create separate bins (for example, 26-27 and 28-30) in the lower level so they could roll up into "young" and "middle," respectively.
Other commonly specified dimension hierarchies are time, geography, product, department, and distribution channel. However, you can create any kind of dimension hierarchy that relates to your business or campaign.
Copyright IBM Corporation 2015. All Rights Reserved.
