Instead of describing data in terms of fields and rows of a table, SiSense Prism uses a Dimensional Model to describe tabular data.

 

A Dimension is a logical entity that provides quick access to one or more values in a field.  These values are held as Members within a Dimension.  For a field containing first names, the ‘First Name’ Dimension would contain a single Member for each unique name within the field itself.  If a value appears more than once in a field, a single Member would represent all appearances of this value.  Therefore, if the name ‘John’ appears several times, the ‘First Name’ dimension would still contain exactly one ‘John’ Member.

 

A Measure is an entity that represents a calculation over values in a field.  Apart from the field it is bound to, a Measure is defined by an aggregation.  The aggregation defines the formula to use in order to calculate the final value.

 

 

 

 

There are 6 types of aggregations:

Aggregation	Description	Applicable Fields
Sum	Calculates the total of values	Fields containing numeric values
Avg	Calculates the average value	Fields containing numeric values
Min	Calculates the minimum value	Fields containing numeric values
Max	Calculates the maximum value	Fields containing numeric values
Count	Counts the number of unique values	All fields
Count Duplicates	Counts the number of rows	All fields

 

Let’s take a look at simple example.  Consider the following tabular data:

First Name	Last Name	Month	Work Hours
John	Smith	January	120
John	Williams	January	130
John	Smith	February	110
John	Williams	February	135
Mary	Jones	February	140

 

Possible Dimensions for this data are:

- First Name: containing 2 Members, John and Mary

- Last Name: containing 3 Members, Smith, Williams and Jones

- Month: containing 2 Members, January and February

 

Possible Measures for this data are:

- Total Work Hours: calculates the total work hours by applying the Sum aggregation on the values in the Work Hours field, resulting in 120+130+110+135+140=635

- Average Work Hours: calculates the average amount of work hours by applying the Average aggregation on the values in the Work Hours field, resulting in (120+130+110+135+140)/5=127

- Number of Work Months: calculates the number of different month values in the Month field, resulting in 2 (January and February).

The most powerful thing about Dimensions and Measures is that they can be easily combined to perform calculations over isolated data at any granularity.  Combining the ‘Month’ Dimension with the ‘Average Work Hours’ Measure would calculate the average work hours per month.  Similarly, combining the ‘Last Name’ Dimension with the ‘Number of Work Months’ Measure would calculate the number of different months each ‘Last Name’ worked.