Natural Objects

Natural objects are Ursa Studio objects designed to capture the basic facts of a concept in a standardized format, without any unnecessary interpretation or presumption of what purpose the data will eventually be put to. That is, they are both source-agnostic and application-agnostic.

Natural objects are generally recognizable as the most fundamental building blocks of more complicated analyses -- the concepts often represented by an enterprise data warehouse's fact and dimension tables. For example, the Natural objects in the Ursa Health Core Data Model include Patients, Providers, Pharmacy Claims, Payors, Medication Orders, and Appointments.

The Natural Object Layer refers to the the collection of all Natural objects in the data model; similarly, it is also used to refer to the stage of the data journey in which data are stored in Natural objects.

Under normal conventions the Natural Object Later is the earliest stage in the data journey in which data from all source systems are comingled. Natural Objects can therefore be usefully thought of as the dividing line between the source-specific objects used in the integration of new data sources and the source-agnostic, analytics-ready objects. That is, Ursa Studio users engaged in the integration of new data sources should push data to the Natural Object Layer and no further, and users engaged in the production of analytic deliverables should not reach back further than the Natural Object Layer.

FAQ

• Q: What is the significance of the word "natural" here?
  A: The word "natural" is meant to convey that the data contained by these objects are presented without interpretation, in what might be considered their natural state. A word like "raw" would not be appropriate, because the data have likely undergone extensive cleaning, mastering, and reshaping to conform to a source-agnostic standard by the time they arrive in the Natural Object Layer. Another contrast is with the word "synthetic", which is used to describe downstream objects that are not generated directly from source data but rather have been created by interpreting and/or combining one or more Natural objects or other Synthetic objects.