Solving the Learning Data Puzzle: The Unizin Common Data Model

Published September 13, 2017

If you’re thinking about how to improve student success – and who isn’t in higher ed? – then data and analytics about teaching and learning are probably on your mind. It can be a daunting subject. There is so much data all around us – about students and courses; the planning, design, and delivery of instruction; and learning behaviors and results – that it may feel overwhelming. How can all of this data be brought together and made useful for faculty, instructional designers, researchers, other staff, and even students? The Unizin Common Data Model (UCDM) is the Unizin Consortium’s approach to solving the teaching and learning data proliferation puzzle. We will explore why this puzzle is so hard to solve, why the UCDM is a sound consortium solution, and how the UCDM will work.

A difficult puzzle for higher ed

It’s not as if data isn’t already in use across the Academy. Many data sources are being used to support the business needs of education. But teaching and learning poses a new set of data challenges: massively increasing diversity of data and its sources. As the shift to digital mediation of pedagogy advances, more of the learning process can be measured directly. At the same time, the ecosystem of content, tools, and learning environments is expanding. The next generation digital learning environment (NGDLE) is a blueprint that promotes this diversity. From student information systems (SIS) to learning management systems (LMS) to learning tools, assessment tools, and content delivery systems, the number of silos where teaching and learning data is captured and stored continues to trend upward. Some of the information is fairly static, like student demographic, socioeconomic, and pre-enrollment data. Other information changes at the pace of the curriculum, term by term, such as courses, enrollments, grades, and majors. We call these “slow” and “medium” data, respectively. Taken alone, they are not particularly challenging to manage. “Fast” data is information about instruction and learner behavior that is generated as it occurs. The learning environment and tools can store this data in their silos, and often do, slowing the data down. As real-time information becomes more a part of everyday life (think of Twitter, Facebook, and Uber), our expectations are that “fast” data should be put to use immediately to notify and inform those with a role in student success. A solution is needed to connect the dots between foundational, structural, and behavioral information to create a single dynamic view of a student in the context of learning. These data elements are of little value alone; it is only when they are connected that they reveal actionable insights. As the number of separate data stores grows, so does the complexity of finding the information one needs to answer key questions. With the pieces so scattered and isolated from each other, this is a difficult puzzle to solve.

Why is the UCDM a sound consortium solution?

Members of the Unizin Consortium realize the need to take ownership of this data proliferation puzzle. Ecosystem diversity addresses pedagogical and business needs, but it introduces challenges to creating a unified picture of the student. The solution must anticipate these challenges. The Unizin Common Data Model organizes a standardized data repository for each Member by combining the silos of teaching and learning data into a set of managed and extensible  resources. Doing this once as a consortium yields five key benefits:

1. Cost – One shared solution reduces the implementation cost for everyone. Operational costs (engineers) are lower because they are spread across the consortium. Infrastructure costs (servers, networks, and storage) are less expensive when demand is aggregated.

2. Focus – Consortium members can concentrate more on research, insights, and action and less on wrangling data.

3. Standards – Industry standards (CEDS, Caliper) help to expand the ecosystem by making it easier to get data in and out of the Unizin Data Platform and the UCDM.

4. Common – A common data repository approach allows reusable solutions across the consortium. Reports, applications, analytics models, and data feeds using UCDM-based resources can be innovated upon by Member institutions and easily shared with others.

5. Research – The common data model allows Members to opt-in to a large-scale, de-identified, anonymized, and aggregated data resource for longitudinal, cross-institutional teaching and learning research. This unprecedented collection of curated data across multiple institutions will be called the Unizin Learning Laboratory.

How will the UCDM work?

UCDM is a schema that governs how ingested data will be organized into resources. It is at the core of the Unizin Data Platform (UDP). The UDP architecture handles acquiring all of the data – slow, medium, and fast – and organizing it according to the rules of the UCDM. The UCDM rules map student, course, instruction, and learning activity information together. They solve the problem of “connecting the dots” between all of the data sources to create a single view of the student in the context of learning. As the data flows in from the SIS, LMS, and learning tools, the rules are applied to each data element, like a puzzle piece, to make sure that it is oriented to contribute to the whole picture. Two industry standards guide these rules and shape the big picture: CEDS and Caliper. While a deep dive into either standard is beyond the scope of this post, let’s dig in just a bit to understand how each standard contributes to the UCDM. The Common Education Data Standards (CEDS) project is a national collaborative effort to develop voluntary, common data standards for a key set of education data elements to streamline the exchange, comparison, and understanding of data within and across P-20W institutions and sectors. The National Center for Educational Statistics facilitates the CEDS project. Most slow to medium data is well-modeled in CEDS, but since our goal is to assemble as complete a picture as possible, the UCDM is actually a bit more than CEDS. We call it CEDS++.

• The first “+”  is for extensions to bring fast data into the picture using the Caliper Framework from IMS Global Learning Consortium. Caliper is the emerging fast data standard for education. It is a means for consistently capturing and presenting measures of learning activity, by far the largest number of pieces in the data proliferation puzzle.

• The second “+” is for extensions that are unique to each Member institution. Unique vocabularies and data elements can be added to each Member’s UCDM store when a consortium-level solution is not feasible due to local needs.

The Unizin Common Data Model is a key component to solving the biggest teaching and learning questions by placing each data puzzle piece into a unified view of the student’s digital higher education experience.