Comparison of Aggregation Types

In LityxIQ, two different types of data aggregations are available, which can be used in conjunction with each other.  The two types can be described as:

1) aggregating over all combinations of the levels of the selected variables,.

2) aggregating over all individual levels of the selected variables.

This document provides a comparison of these two techniques.  Both area available in the Aggregation area of defining a View.

Assume we have a very small and simple dataset with three variables: Age, State, and Gender.

Age State Gender
30 MD M
30 MD M
35 MD F
40 MD F
30 VA M
35 VA F
45 VA F
45 VA M
50 VA M

There are a variety of ways we can aggregate this dataset using the Combinations method.  One, two, or three variables can be used to aggregate the dataset according to this first method.  Some examples are below (note: in all of these examples, we are assuming that the only summary statistic we are computing is the Record Count):

We also have options to aggregate this dataset using individual variable levels.  One, two, or all three variables can be selected.  Here are some example outputs:

Notice that the individual variables option provide an output that include the Variables "_Variable" and "_Value".  In combination, these show for each row which variable and which variable value are represented by the other data in that row (in this case, just the Count variable).  Another thing to notice is that these individual variable results are essentially just the results from having run the variable combinations method over and over again for one variable at a time.

Now, there is also the option of using both of these methods together.  Below are two examples of this.  Essentially, the Individual Variables concept is applied one at a time to each combination of the Combination Variables selected.