Data models
Last updated
Last updated
The UDP integrates context data and behavior data from different systems and tools. The UDP requires context and behavior data to conform to standard data models. When necessary, the UDP will transform learning data into the relevant standard data model. Consequently, learning data sourced from different learning tools will share a single data model. By combining the data from different systems and tools, a UDP instance creates a more comprehensive picture of learning, learners, and learning environments.
Transforming data into a single data model is a key step in combining data from different tools into an easy to understand whole together. Additionally, transforming the data into a single data model makes downstream data use by data scientists, researchers, developers, and other end users more convenient and straightforward.
The UDP uses two standard data models: the Unizin Common Data model for context data; and the IMS Global Caliper model for behavior data.
The Unizin Common Data Model, or UCDM, is a data interoperability standard designed to exchange context data between learning tools and the Unizin Data Platform. The UDP accepts context data only if it conforms with the UCDM. Any incoming context data that does not conform with the UCDM will be rejected by the UDP (see the concept of a UDP loading schema for more details).
Because context data is relational, the UCDM is also a relational schema that can be used to instantiate a datastore to support UCDM-conformant data. The UDP's Context store, for example, is a physical instantiation of the UCDM's normalized, relational model.
The Unizin consortium shepherds the development of the UCDM. It invites the participation of its members and vendor partners to develop the UCDM.
The Unizin Common Data Model is defined by key concepts that enable it to describe learners and learning environments. These concepts are called entities, elements, and option sets.
An Entity is a high-level concept in the UCDM that refers to a first-class object in the UCDM ontology. Such first-class objects include, for example, Person, Academic term, Academic degree, Learner activity, and other core elements of the data ontology.
The UCDM can be instantiated in a database system as a relational data model. Consequently, Entities have relationships to each other as defined by their foreign key relationships. An ability to preserve relationships between the entities that define learners and learning environments is key to creating value with learning data sourced from an institution’s tool ecosystem.
A Option set is an enumerated set of allowed values for an Element.
In the UCDM, some elements may only take values from a finite set of possible values that are valid for the element. For example, the Person element that describes an individual’s sex may only take the values “Male,” “Female,” NotSelected,” and “NoData.” All other values are invalid and do not conform with the UCDM.
The UCDM defines a wide variety of option sets in its data model. Context data integrated into the UDP must align its values with the enumerated option set where they exist.
The UDP accepts behavior or behavior data if it conforms with the IMS Global Caliper standard, whose data model is often referred to simply as “Caliper.” While the UCDM describes static objects, The Caliper data model describes discrete behaviors/events in the learning environment.
As with the UCDM, the Caliper data model defines an ontology to which data generated by a learning tool must conform. That ontology is maintained by IMS Global’s Caliper working group. However, many aspects of its ontology (nouns and verbs) align closely with those of the UCDM.
