Base Dimensions

Dimensions are the lists of keys that are used to uniquely identify a data point.

Example:  If I tell you that card ID 1234567 purchased 4 bottles of item 3500012345 at store number 345 on June 15, 2015 at 3:15 in the afternoon and paid a total of $17.00 for those items that usually cost $5.00 each, I am identifying the data using these dimensions


Store Number 345 – Usually thought of by users as Store, or Location.


item 3500012345 – This is the Item dimension.  Values are some form of UPC/PLU


June 15, 2015 at 3:15 in the afternoon – The transaction data is collected with a TIMESTAMP which is time down to the second.


card ID 123456 – The most granular level of shopper is the loyalty card ID.

The dimensions above are used to collect the data and are found in the transaction log file generated by the Point Of Sale.

Other Dimensions / Attributes

There are other lists of keys that we employ, but instead of identifying data points directly they are used to identify other dimension values.  We’re calling those Dimension Attributes, they relate to a base dimension, and they are things like

  • Brand, Manufacturer, Size, Diet – These are Product Attributes used to identify groups of items
  • Division, Group, Area, Format – These are Geography Attributes used to identify groups of stores
  • Week, Month, Period, Quarter, Year – These are Time Attributes used to identify groups of days
  • Gender, Value Segment, Price Segment – These are Shopper Attributes used to identify groups of shoppers.

Aggregation Schemes / Hierarchies

Often we will want to present data at an aggregate level, Total for All Stores is an example.  Even more often we find it handy to deliver results at many levels of aggregation in an overall aggregation scheme, called a drilldown hierarchy.  The word “drilldown” refers to the ability of the user to expose successively lower levels of detail of a hierarchy node by clicking on the row heading and being shown the child nodes of the node s/he clicked on.

Here are some sample hierarchies and the dimensions to which they apply.

  • Notice how you can have more than just one aggregation scheme identified for a  dimension.
  • Notice how the levels of aggregation are controlled by an Attribute of that Dimension.


  • Store >> Division >> Region >> Total
  • Store >> Format >> Region >> Total
  • Store >> HasFuel >> Region >> Total
  • Store >> HasFuel >> Total


  • Item >> SubCategory >> Category >> SubDepartment >> Department >> Total
  • Item >> Brand >> Category >> Total
  • Item >> Brand >> Manufacturer >> Total
  • Item >> Brand >> Total


  • Day >> Week >> Quarter >> Year
  • Day >> Week >> Period >> Year
  • Day >> Month >> Quarter >> Year
  • Hour >> DayPart  >> Day


  • CardId >> HouseholdId
  • HouseholdId >> Value Segment >> Total
  • HouseholdId>> Value Segment
  • HouseholdId >> Price Segment