Creating a dataset

Applies to v4.5 and later.

1. Attributes (blue) - Descriptive text or time/date series information etc.
2. Measures (red) - Derived numerical information i.e. sum(Sales), count(Employees) etc.
3. Pivots (green) - A Measure dynamically Pivoted by an Attribute.

Note: For more information see understanding attribute and measure columns.

To create a new dataset:

1. Select Main Menu -> Dataset Builder -> New Dataset.
   
   
   
2. Specify the datafeed you wish to base the dataset on by browsing the folder structure or simply typing the title of the connection in the search field.
   
3. You can now choose columns from the datafeed on the left to begin populating the dataset on the right. You can add new dataset columns by dragging-and-dropping the column header from the datafeed on the left to the dataset on the right.
   
4. You can display the datafeed columns as a list by clicking on the relevant symbol on the toolbar: 
   
5. Now you can easily search for specific datafeed columns by typing a search string.
   
   Columns dragged to the dataset are set to be Attributes (blue) by default. You can edit the properties of each column in the dataset by clicking on the column header. The column type can be changed from Attribute to Measure (red) here.
6. Clicking on a column header in the dataset you will be presented with the Column Property dialogue box where you can change:
   -  Name - Display name for dashboards i.e. "Total Pages"
   -  Category - Change the column type here. Attribute or Measure
   -  Type - Change the data type here. String/DateTime/Decimal/Integer
   -  Description - Text description of the column
   -  Default Width - Default column width when displayed in a component
   -  Allow column to be filtered - Allow component filtering Y/N
   -  Expression - Measure expression i.e. Sum({Colour_Pages})+Sum({B_W_Pages})
   -  Data Format - Display formats eg.  £3,120.25 (£ symbol at start, 2dp, show thousand separator) 
   -  Colour Range - Default conditional formatting for measures.
   
7. Columns containing values that are descriptive and will be displayed in a component without having any calculation applied, should remain defined as Attribute columns (blue). i.e. text descriptions, dates, periods etc. These generally form the X-Axis of a chart.
8. Columns containing values that will be used to calculate a result within a component should be changed to a Measure column (red). Once this option is chosen, you will be prompted to choose a function to apply to the data in that column. The functions available are Sum(), Min(), Max(), Count() and CountDistinct(). These functions work by considering many rows within the dataset then calculating a single result. They are named, and calculated, in exactly the same way as the standard SQL aggregate functions found within any relational database product eg SQL Server or Oracle. An overview of aggregate functions can be found here. These generally form the Y-Axis of a chart.
9. Measure columns can contain sophisticated mathematical expressions accessed via the expression Icon.
   
10. Pivot Measures may also be defined here eg. Pivot Total Pages by Printer Type
    
11. All three dataset column types are shown here: Attribute, Measure, Pivot
    
12. It is also possible to create different DateTime columns using a single DateTime object from the datafeed by formatting the DateTime column ie. Year, Month, Day, Period etc.
13. Once you have fully defined your dataset save it using the save icon.
14. Optionally, publish the dataset using using the publish icon.