HOW TO: Aggregated Table vs. Flat Aggregate

HOW TO: Aggregated Table vs. Flat Aggregate

Overview

Aggregates can provide useful calculations and data summarizations of one or many variables. If you do not need to maintain the structure nor all of the details of your dataset, an Aggregated Table can provide a condensed, summarized view of your data. If you do need to maintain the structure and details of your dataset, a Flat Aggregate can provide calculations in a new column without removing any of the contents of your dataset.

Aggregated Table

There are multiple ways to initiate an Aggregated Table, all of which use the Aggregate Transform.

Method 1: Predictive Interaction

  1. Select the Column Header of the column you wish to Aggregate
  2. Select the Aggregate suggestion
  3. Click Edit
  4. Edit the Functions you'd like to include in your table
  5. OPTIONAL: Choose Columns to Group by, which determines which unique values to perform the Functions on

Method 2: Column Menus

  1. Select the Column Menu dropdown for the you wish to Aggregate
  2. Select Aggregate
  3. Choose As New Table
  4. Select the function you would like to use in your Aggregate
  5. Edit the Functions you wish to perform
  6. Choose your Group by variables

Method 3: New Transformation Step

  1. Select the '+' button in the bottom right of the Transformer Grid
  2. Type or Select the Aggregate Trnasform
  3. Fill in the Function and Group by parameters

Flat Aggregate

You can initiate a Flat Aggregate step in the same ways as an Aggregated Table, so for simplicity, we will use Method 3 from above, this time using the Derive Transform. NOTE: You can only perform one formula per Derive Transform.

  1. Select the '+' button in the bottom right
  2. Type or Select the Derive Transform
  3. Fill in the Formula and Optional Group by parameters
  4. You can also select an Order by parameter to sort your data

NOTE: If you are working in a sample, the Aggregated numbers will be representative of that sample, and not your dataset as whole until you run the job/generate results and apply the steps to your entire dataset.