In attendance:
Jim Coleman, Provost and Executive Vice Chancellor
Bob Shea, Vice Chancellor for Finance and Administration and Chief Financial Officer
Tina McEntire, Vice Chancellor for Enrollment Management
Terri Shelton, Vice Chancellor for Research and Engagement
Donna Heath, Vice Chancellor for Information Technology Services and Chief Information Officer
Kelly Burke, Vice Provost and Dean of the Graduate School
Andrew Hamilton, Associate Vice Provost for Student Success and Dean of Undergraduate Studies
Johnny Lail, Business Intelligence Data Analyst, Institutional Research
Larry Mayes, Associate Vice Provost and Director of Institutional Research
Lee Norris, Associate Vice Chancellor, Data Architecture, Reporting and Visualization
Chris Partridge, University Registrar
Jodi Pettazoni, Associate Vice Provost and Director of Assessment, Accreditation and Academic Program Planning
Topic: Institutional Data Governance Overview
Topic: Data Lake and Data Presentation Layer Overview
Topic: Identifying short-term data and reporting needs
Topic: Building a long-term data management, reporting, visualization and analytics strategy
Context: UNCG leadership does not currently have access to the operational and management data required for effective decision making. Decisions about how data are coded and reported have occurred historically in a siloed fashion within work units and these stand-alone decisions have had a negative impact on the overall quality of our data at the institutional level. Ad hoc reports and dashboards have been developed within functional units that leverage data elements inconsistently and result in disjointed reporting and interpretation of data across groups.
Our Banner Enterprise Resource Planning (ERP) environment is complex. It consists of multiple core applications (Finance, Student, Financial Aid, HR/Payroll, Advancement, Research and Sponsored Programs, etc.) and many ancillary systems. The Banner application suite contains more than 5,000 data fields in 5,126 data tables and 2,126 data views. There are more than 60 ancillary systems connected to Banner that consist of thousands more tables and views. In addition, we have access to IPEDS tables, the various datasets created in IR, Delaware Survey data, and a number of System Office datasets. The 60+ ancillary applications contain rich data about our students, faculty and staff that we don’t leverage today as part of the institutional decision-making process because these applications exist siloed in functional workgroups.
A single data element can appear in multiple places and can have different meanings depending on context. Credit hours data in Banner, for example, could mean “enrolled credit hours” or “billable credit hours” depending on context, whereas “fundable credit hours” does not appear in Banner at all but instead is constructed based on State rules, and is stored in the UNC Student Data Mart tables. In each of these instances, the credit hours numbers will be different. Further, data in Banner and the ancillary system tables change throughout the day as functional staff execute transactions in the course of their work. Data reported from live tables can change from minute to minute. Understanding these data elements and using them correctly requires advanced expertise in both the data and the systems.
To further complicate things, we leverage multiple (and often expensive) tools across the institution to accomplish the same functions depending on the preference of workgroups. For example, we have both the WebFocus and Argos enterprise reporting suites at an annual cost to the university of $40K and $41K, respectively. Many of these tools deliver essentially duplicate functionality. Our discussion will focus on action items to address the immediate operational management data shortage as well development of a longer-term data management, reporting, visualization and analytics strategy.
Discussion continued to 8/20 and 9/3.