Data Stewardship I

What is the definition of Data Stewardship (UNECE)?

Data stewardship is concerned with the science and practice of data collection for the purposes of analysis, reflecting the values of fair information practices. In practice it is a collection of methods and mechanisms of data management encompassing acquisition, storage, protection, aggregation, deidentification, and procedures for data release, use and re-use — to ensure that data assets are of high quality, easily accessible, and used appropriately.

What is Data Governance (Plotkin)?

Data governance concerns decision-making and authority for data-related matters, whether within or between enterprises and public agencies. It is a system of decision rights and accountabilities for information-related processes, executed according to agreed-upon models governing the kind of data stored, the authority to access data, and the methods of data access.

What is the key difference between Data Stewardship and Data Governance?

Data Governance is about HOW people manage and make decisions about data — it is about the decision-making framework. Data Stewardship is about ensuring that people are properly organized and do the right things to make their data understood, trusted, of high quality, and ultimately suitable and usable for the enterprise's purposes.

What are the 5 responsibilities of a Data Steward?

1) Ensuring data quality, data definition and privacy standards are met.

(2) Ensuring that data is fit for purpose (completeness, accuracy and integrity).

(3) Managing metadata and processes to ensure proper use of data being read, created, collected, reported, updated or deleted.

(4) Ensuring data is protected and security procedures are enforced.

(5) Taking an active part in the data governance framework to feedback on existing practice and recommend improvements.

What are the 3 key competencies of a Data Steward?

Departmental knowledge — deep knowledge of the operational area including processes, rules, data flows and data sources.

Communication skills — ability to interpret and communicate policy or business rules to end users, and to feed ideas back to technology and policy owners.

Collaboration skills — working with other Data Stewards and stakeholders across the organization to ensure data flows smoothly.

How is a Data Governance Program typically structured?

As a pyramid, with support from IT and a Data Governance Program Office.

From top to bottom:

Executive Steering Committee → Data Governance Board (Data Governors) → Data Stewardship Council (Business Data Stewards) → IT Support.

What is the role of the Executive Steering Committee?

It is authorized to change the organization, drives cultural change, supports the program enterprise-wide, and provides funding for the Data Governance Program.

What are Data Governors and what do they do?

Data Governors are high-ranking representatives of data-owning business functions who can make decisions about data for the company. They assign members of the Data Stewardship Council, approve decisions of the Council, and approve data-related policies.

What are Business Data Stewards and what do they do?

Business Data Stewards are experts on the use of their data. They reach out to SMEs (Subject Matter Experts) to gather information and make decisions, know who their stakeholders are, and are typically the most knowledgeable about the meaning of the data. They make recommendations on data decisions, write data-related procedures, and may be part of a data domain steward team.

What is the Business / IT Alignment in Data Governance?

The Data Governance Organization has both a Business side and an IT side. On the Business side: Data Governance Business Sponsor, Data Governance Manager, Data Governors, Enterprise Data Steward, Business Data Stewards (PT), and Project Data Stewards (FT). On the IT side: Data Governance IT Sponsor, Enterprise Application Owner (PT), and Technical Data Stewards (PT). PT = part time, FT = full time.

What is the "Information Chain" in Data Stewardship?

Information flows through an enterprise in an "Information Chain" — data moves from its source through various processes of collection, transformation, storage and distribution, until it reaches its final users. The Data Steward's role is to ensure this chain runs smoothly, with data remaining trustworthy, understandable and of high quality throughout.

How many G's are in Applied Data Stewardship and what are they?

There are 4 G's (according to Rancourt, 2019): Guard, Gather, Grow, and Give. Note — the mock exam uses 4 G's, not 5. These represent the 4 phases of the data lifecycle in applied stewardship.

What does "Guard" mean in the 4 G's?

Guard means paying special attention to access rights and privileges, performing data audit trails, systematizing data monitoring and back-up protocols, and consistently updating metadata standards and classification systems. The goal is to adhere to "privacy by design" principles — ensuring data is secure, encrypted, confidential and de-identified, with all necessary privacy protocols in place to function ethically and according to the trust framework.

What does "Gather" mean in the 4 G's?

Gather refers to all data ingestion — collecting and integrating data assets through various systems of acquisition, as well as the policy instruments and ethics-based legislative frameworks through which the agency gains access to data and information. Sound data stewardship ensures that data is acquired efficiently, ethically, and without duplication or redundancy.

What does "Grow" mean in the 4 G's?

Grow means the data is organized, processed, transformed, integrated and extracted for various uses. During this phase, data is cleaned and verified, quality assurance is performed, data is analyzed, explanations are developed, and hypotheses are tested. Efforts are made to optimize data and adhere to data quality frameworks.

What does "Give" mean in the 4 G's?

Give means data and statistics are shared and published. Data access and interoperability are ensured, dissemination occurs regularly with quality and accessibility, and appropriate metadata is made available. The goal is to increase data discoverability and be "open by design" — with sharable and open data, metadata, metainformation and analysis.

What is a DMP / DMSP and why is it important?

A Data Management Plan (DMP) or Data Management and Sharing Plan (DMSP) is a key element of good data management. It describes the data management life cycle for data to be collected, processed and/or generated in a project (e.g. Horizon 2020). Some funding agencies have their own DMP policy. Tools like DMPTool can be used to create one.

What are the 6 required elements of a DMSP?

(1) Data Type — identifying the estimated type and amount of data to be generated, which data will be preserved and shared, and the accompanying metadata.

(2) Related Tools, Software and/or Code — tools needed to access and manipulate data.

(3) Standards — standards to be applied to scientific data and metadata.

(4) Data Preservation, Access, and Associated Timelines — proposed repository, how data will be findable, when it will be available and for how long.

(5) Access, Distribution, or Reuse Considerations — informed consent factors, privacy protections, whether access to human data will be controlled.

(6) Oversight of Data Management and Sharing — who will monitor and manage plan compliance.

What does the "Data Type" element of a DMSP specifically cover?

It covers: identifying the estimated type and amount of data to be generated (modality, level of aggregation, degree of processing), which data will be preserved and shared, and the accompanying metadata, other relevant data and associated documentation to be made available. This is the exam-relevant element for the DMSP question in the mock exam.

What are the key takeaways of Data Stewardship I

ata management is important to keep data safe from harm and make data usable and discoverable. A data management plan includes strategies and processes to organize, describe, preserve and share data. Some funding agencies have their own DMP policy. DMPTool can be used for templates and guidance. Data Stewards do NOT work individually and manually — they collaborate with others across the organization.