U-M Data Governance Revitalization

In order to take advantage of big data and stay at the pinnacle of data science, U-M must update its data governance and stewardship model, which was created in 1994 and not substantially updated since. To do so, representatives from many university areas are working to identify and, where possible, quickly implement improvements to data governance processes.

This revitalization effort is jointly sponsored by the Provost and the VPIT-CIO.

We are approaching the revitalization effort in multiple stages.

Stage One: Create High-level Recommendations (Complete)

The high-level recommendations were published in 2019. They are:

  • Create a Data Governance Oversight Committee (DGOC) Structure
  • Provide dedicated resource support
  • Create mechanisms to provide systematic logging of all requests for access to data
  • Define and provide central services that can support and guide any data request
  • Reduce the number of, and better define the roles within, the data governance framework

Details for Stage One

Stage Two: Recommend Actionable Improvements (In Progress)

A Data Governance Council was established in April 2020 to work on Stage Two of data governance revitalization. The council has identified the following areas of focus for their work in 2020:

  • Data Availability and Sharing
    Document existing and desired processes. Establish common guidelines for obtaining access to data. Wherever possible, automate data request processing.
  • Data Quality
    Identify opportunities for improvement in how data quality issues are addressed. Work to improve consistency of data definitions. Publish data quality metrics.
  • Roles, Responsibilities, and Appointments
    Simplify data stewardship structure. Create consistency in roles and alignment on responsibilities. Improve maintenance of Data Governance structure.

Details for Stage Two

Data Catalog Implementation

U-M does not currently have a data catalog that covers data sets and their metadata across the university. Such a data catalog will be instrumental in successfully achieving several of the high-level objectives from Stage One. We are currently exploring a data catalog that is included in a tool ITS uses for Data Virtualization, from Denodo.

Questions

Please contact the Data Governance Council co-chairs with any questions you may have at dg-chairs@umich.edu