Data Definition

Prior mandatory steps 1) Upload dataset 2) Create analysis 3) Create sub analysis

A datafile typically contains several attributes (columns). An analysis designates one attribute as the target. If you want to fine tune remaining attributes, use the Data definition API. For example, some attributed may be actionable in the sense that their value can be changed. , discount_type may be changed while customer_age is not. Such attribute level settings happen during this stage.

Attributes may be 1) Renamed by setting the business name 2) Ignored from analysis by moving them to “IGNORE” group 3) Grouped by creating custom groups. Grouping is a mechanism to organize large number of attributes into meaningful subsets. 4) Some attributes may be set to actionable. 5) May be set to categorical even if they look like numerical (eg ranks 1,2,3,4).