Data categories

The Unizin Data Platform (UDP) imports and manages two categories of data: context data and behavior data.

Context data

Context data describe objects (e.g., learners, assignments, modules, outcomes, learning design, course catalog, degree) and relationships relevant to learners, learning environments, and overall academic experience. Context data exists in all tools and systems used by an institution to support its academic mission.

Generally speaking, context data originates in the operational data stores (ODS) of a learning tool. Context data is usually extracted from an ODS as flat files or other data formats that are suitable for batch processing.

Context data is relational in nature. Its purpose is to describe the relationships between concepts and the concepts themselves. For example, the concept of a Person has many descriptive attributes in the Unizin Common Data Model and, also, relationships to other concepts, such as Course section enrollments (expressing courses taken), Academic degrees (expressing degrees earned), and Learning activities (that are completed). By modeling contextual data relationally, the UDP enables querying patterns that are natural to many business intelligence and analytics contexts.

Institutions that use the UDP will rely heavily on context data. In particular, institutions will need to assure that context data from the Student Information System (SIS), Learning Management System (LMS), and the tools plugged into the LMS are integrated into their UDP instances.

Behavior data

Behavior data describe the discrete actions of instructors, learners, teaching assistants, and even tools themselves in the learning environment. Each discrete action describes a pedagogically-relevant event. Behavior data is typically generated and ingested as a stream of events.

Behavior data is usually emitted by learning tools in real-time. As students engage in their online courses with learning tools, their behaviors are captured by instrumentation, represented in a standard data model, and transmitted to the Unizin Data Platform.

Unlike context data, behavior data is usually not rich in its descriptive elements. Its data payloads are relatively small and focus only on asserting that behaviors happened. They are not rich in the contextual data that contributes to an interpretation of behaviors and activities.

It is sometimes also called "event data" and "process data."