Download presentation
Presentation is loading. Please wait.
1
Data Management Concepts and Data Profiling
WORKSHOP Data Management Concepts and Data Profiling
2
Getting in touch with the STUDENT EXPERIENCE
3
Workshop Agenda Analytics vs. BI Why Data Quality?
1:30 Why Data Quality? :15 Data Validation – When and Why? :30 Data Profiling Concepts – Why and How? 1:00 Data Governance Best Practices Golden Source vs. Master Data Business Intelligence, BPM, & Analytics Data Integration, Enterprise Data Warehouse Data Governance Master Data Management
4
UF Data Management Pillars
Data Governance Data Architecture Master Data Management Data Warehouse/ Data Integration Business Intelligence / BPM / Analytics Data Quality Meta Data Management SOA / ESB / API Management
5
Who Needs What Information?
Bursar’s Office Deans & Dept Heads Office of Student Financial Affairs Office of the University Registrar Office of Research Revenue & Expense Analysis Tuition & Financial Aid indicators Capital Projects Indicators Enrollment, Instructor, & Course Offering metrics Faculty hiring and retention metrics Research Analytics Student Progress Reporting Instructor Course Workload Program Effectiveness metrics Graduation Rates & Time-to-Degree metrics Student Holds Dean’s List Board of Governors Compliance Reports Disbursement Reports Board of Trustees Reports Fiscal Operations Reports & Indicators External Agency Reporting Ad Hoc Reports Institutional Reporting Student Holds Student Fees & Payments Financial Operations Reports Financial Operations Metric & indicators Loan Reconciliation Reports Cash Account Reporting Grant Reconciliation reports Student Holds Student Enrollment Registration metrics Students with Repeat Courses Board of Trustees Compliance Reports Federal & State Compliance Reports Official Student Metrics Withdrawals – alerts, monitoring Student Holds Official Institutional metrics reporting Grant Funding Grants Management – alerts & monitoring Grants Lifecycle Dashboard Ad Hoc Reporting Common Data Sets Reporting Academic Leadership Fiscal Officers Dept Admins Managers Principal Investigators Grant Administrators
6
Where is My Information?
Finance HR Payroll Research Accounting Budgeting Purchasing Admissions Financial Aid Student Records Scheduling Registration Accounts Receivable Housing Institutional Research, Health Services, Sponsored Research, Enterprise Infrastructure & Operations, Environmental Health & Safety, University Relations Student Portal UF Online Registration
7
What Do We Need to Know? Profile of Students most likely to succeed
Faculty Productivity: Research, Publications, Teaching, Advising, etc. Affordability and Financial Aid: socioeconomic profile of our student body Early detection of students struggling academically => early intervention; Advising Cost/Benefits Analysis and Assessment of Academic Programs Per Student Cost to attract, retain, and graduate Supply and Demand of course offerings Which Alumni are likely to make large donations Research: Proposal success rates, growing areas, funding opportunities, Cost Sharing and Indirect Cost projections/commitments Libraries: Subscriptions Cost/Benefits and Allocation models Cost Reduction-Financial and Operational indicators: identifying inefficiencies and duplications, non- value added functions Space utilization and optimization Adapted from "Practical Approach to Implementing Business Intelligence in Higher Education," Ora Fish, Executive Director, Program Services Office, New York University Financial Aid Revenues
8
UF Information Factory
Data is an institution asset and must be managed accordingly BI & Analytics Data is a valuable institution resource; it has real, measurable value. Data is used to keep accurate records of operations and aid in operational and strategic decision-making. Accurate, timely data is critical to accurate, timely decisions. Most institution assets are carefully managed, and data is no exception. Discovery Analysis Data Governance UF Information Factory Master Data Management Data Integration (warehouse) Data Quality Data Profiling
9
Building Blocks for Data Management
Business Intelligence Analytics Business Performance Building Blocks to achieving a HOLISTIC and INTEGRATED view of the Student Each Building Block is required to achieve TRUSTED and QUALITY information SPONSORSHIP FACULTY STUDENT SCHOOL Data Integration Master Data Management Best Practices : Set standards and approach Choose right tools Build a customer-centric, Enterprise Data Warehouse Data Quality Framework Data Governance
10
UF Data Principles – The Foundation
Aspirational Description Primacy of Principles These principles apply to all faculty, staff, functions, data, and processes within the institution Executive Mandate Institution is mandated to adopt a data culture and motivate people to apply data principles Our Data is an Asset Data is an institution asset and must be managed accordingly Our Data is Shared Data is shared across business units in order to avoid duplication and enable reuse Our Data is Accessible Enterprise data assets are available to business users when and where they need it within predefined business constraints Our Data is a Manufactured Product Data is a product made from the large-scale operation of business processes Our Data is Governed Enterprise data assets require specific ownership, stewardship, governance and quality controls Our Data is Meaningful Enterprise data is defined consistently using a common vocabulary that is managed and accessible across the institution Our Data Quality Must Continuously Improve Critical data elements are continuously identified, controlled and monitored according to a selected set of quality dimensions Data is Protected Sensitive data is identified and protected according to firm policies, and monitored and controlled for compliance to the policies Our Data Architecture is Change Enabled Our data architecture framework is sufficiently flexible to meet current and future business requirements Our Data has a Lifecycle The institution creates, reads, updates, archives, deletes, and purges data throughout the information lifecycle guided by industry best practices and regulatory requirements Our Data Must be Harmonized The institution guards against distrust, loss of value, and misuse due to misinterpretation and ambiguous meaning Our Data Must be Authoritatively Sourced Data is consumed from the most appropriate, governed sources Data Management Education is Essential As a valuable asset, the institution as a whole has ongoing education and training to manage data in a way that optimizes the value of the asset 9/16/2018
11
Business Intelligence
Analytics vs. BI 1:30 Business Intelligence, BPM, & Analytics Business Intelligence
12
Business Intelligence Overview
Interactive, discovery, investigative analytic approach that recognizes the cycle of analysis: each answer brings new questions Question Scorecards & KPIs Analytics & Reporting Operational Reporting Summary Data Official Revenue Official Student Counts Performance Dashboards & Alerts Answer Question KPI Drill Down Slice and Dice Discovery Analysis Algorithms, trends, predictions Answer Drill-to Detail Operational Alerts & Dashboards Operational Integration Structured, consistent views Question Answer UFIT Enterprise Data Warehouse Atomic Data
13
BI / Analytics Capabilities Maturity Model
Analytic Technique Critical Business Question Optimization Modeling How can we achieve the best outcome? Predictive Modeling What will happen next if ? Forecasting What if these trends continue? Simulation What could happen .. ? Statistical Analysis Why is this happening? Scorecards / KPI Did we meet our goals? Dashboarding (Alerts) What actions are needed? Discovery Analysis (Query / Drill-down ) What exactly is the problem? Ad-Hoc Reporting How many? How often? Where? Standard Reporting What happened? Analytics Predictive and Prescriptive Support new business models and opportunities Competitive Advantage Business Intelligence Discovery and Diagnostics Support ongoing business operations Data Cleansing and Integration (Data Warehouse and MDM)
14
Example Dashboard for Higher Education
KPI’s Score Card
15
Recruiting Metrics – Example Dashboard
16
What Can BI Do for Higher Education?
"Imagine leveraging the historical data already stored in your data repositories to determine which students are most likely to drop out or not pay their tuition bill on time, which are likely to switch majors or become alumni who generously give back to the campus, or predict where crimes may occur on campus, thus allowing you to staff campus security accordingly.” -- Scott Cupach, Senior Consultant, Sungard Higher Education "Our philosophy with business intelligence is to enable as much transaction-system content as possible to our end users and to empower departments to do a lot of reporting and dash boarding on their own through training” - Michael Barrett, CIO, Florida State Three core principles for successful BI in Higher Education: “ Start simple and evolve, minimize variables, and link insight to action to provide continuous institutional effectiveness.” - Selim Burduroglu, Oracle Education & Research "The BI tool helped us use data and trend analysis to prepare for larger classes and increased enrollment. We also were able to educate stakeholders, like academic affairs, so they could hire to prepare for increased enrollment.” - Denise Groves, Dean of Enrollment, Tarleton State University, Texas 1 2 3
17
Master Data Management
Golden Source vs. Master Data 1:00 Master Data Management
18
Benefits of Data Stewardship
MDM Roles and Benefits Data Trustee Typically core office(s); oversees all business function(s), DDD level, business-driven. Commonly responsible for data content, context, and associated business rules. Sponsor: Champions the MDM program and promotes business and cross-functional participation. The Sponsor empowers the Data Governance Council and the Data Steward; leads local governance and participates on enterprise wide data governance. Data Steward Typically core office; day-to-day functions; adding / editing relevant data, data quality assurance, specific business function(s). Data Custodians Typically IT-focused; can be institutional or local; responsible for the safe custody, transport, storage of the data and implementation of business rules. Data Users / Recipients The person, under direction of their unit VP or Dean, requesting and who will make use of transferred data, typically for reporting or analysis Benefits of Data Stewardship Consistent use of data management resource Easy mapping of data between computer systems and exchange documents It has been found that people are more likely to trust and use a system where there is a person they can call with question on each data element
19
What is Master Data? Places People Things
Data that describes the KEY “people, places, things” (nouns) of the University that need to be managed Data about the people, places, things that will participate in events Provides contextual data about events and transactions People Student Vendor Faculty PI Places address building classroom Sponsor Things Catalog Degree Course Credit codes Reference Data
20
What is Transaction Data?
AKA Event data Describes an action (a verb): e.g. “buy” May include measurements about the action: Quantity bought Amount Paid Includes information identifying the nouns that were involved in the event (the Who, What, Where, When, How and maybe Why): Anna Adams went to the campus bookstore on Friday, December 11, 2015, and bought a book for her Math 101 class. Anna paid $100 cash for the book. Anna Adams Math 101 book Campus Bookstore Friday, December 11, 2015 Cash Does not include information describing the nouns: Anna is female and works for University of Florida IT Friday, December 11th is a school holiday The address for the campus bookstore is 232 Stadium, Gainesville, FL
21
What is Master Data Management?
The processes, systems, and procedures used to manage the relationships between master and reference data PI People Things Reference Data Catalog Student Places Vendor Faculty Sponsor Degree Course Credit address building classroom codes codes
22
UF MDM HUB - Capabilities
Reference Catalog Person 3rd Party Access Data Query Workflow Reporting & Analysis Data Governance Management Performance & Scalability Matching & Merging Real-Time Data Integration Hierarchy Management Enterprise Data Model
23
Business Objectives Establish a single version of the truth for data over time Eliminate unnecessary data duplication and proliferation Take ownership, responsibility and accountability for the improvements of University information quality, accuracy and consistency Implement a flexible and scalable data integration framework of high quality and one that supports agile development
24
Key Drivers for Moving to a Student MDM Hub
Do you have Master Data Management best practices implemented in your OSS? Does each operational domain (School, College) define it’s definition of Student based upon it’s own operational needs, resulting in multiple student definitions across the enterprise? Does each of your OSS systems pass a single student definition/ID across all systems? Are your student IDs linked to the operational process or the actual student? Do you have a single view or 360 of the student for student services or analytics? If not then you might … Not know when the same student enrolls at two different campuses or for credit and non- credit courses Not know when the student has moved off-campus or moved to another dorm Send multiple invoices to the same student or incorrect billing information Send duplicate information to the same student Do you want to send multiple pieces of mail to the same customer Reduce costs for postal requirements – Accuracy of customer information is critical to effective marketing
25
MDM Benefits Improved Recruiting Effectiveness
More efficient Enrollment and Registration process Improved tracking of prospects Improved Student Retention Know past interaction of student with University Offer programs, courses, and scheduling based on student demographics and churn patterns Scalability Integration of disparate information Prospect management, admission applications, transcript evaluation, registration, academic history, student holds, fees, financial aid, contact management Easily support mergers, multi-campus locations, target marketing, academic structure changes Supports CRM and BI Integrate with University’s MDM architecture Master service/product Facilitate 3600 view of student by having one version of the truth DW BI MDM ODS
26
UF Data Management Reference Architecture
Transactional Layer Integration Layer Distribution Layer UF Standard User Enterprise Systems Enterprise Bus. Intelligence Click Commerce D I R E C T BI Platform Meta Data _ Dashboards UF Advanced User People Soft Reporting Scorecards College / 3rd Party E S B Canvas ETL ODS Enterprise Data Warehouse Developer / Advanced Tools DQ Visualization Research Computing Analytics UF Developer A P I Big Data SQL / Direct Analytics MDM Data Steward MDM HUB API MDM Work Flow Data Architecture
27
MDM | Seven Building Blocks of Success
Master Data Management Data Migration and Integration. ETL tools for Person data loads, distribute, replicate and monitor Data quality processes for MDM practices. Profile, Analyze, Cleanse and monitor. Metadata for business and technical documentation Different Person Views catering to Enterprise Wide departments, colleges, and affiliates Enable functionality to create and manage Person Hierarchies at all levels Data Governance and Stewardship. Business Process for data stewardship over hierarchies and user-managed data. MDM Vision MDM Strategy MDM Metrics MDM Governance | MDM Organization MDM Processes MDM Technology Infrastructure Gartner
28
MDM Maturity Model Maturity Stages Level of MDM Maturity Non-existent
MDM is the way we do things around here. Managing master data as an asset. Continuing to learn and improve. A unified vision emerges with high-level sponsorship. Enterprise-wide MDM program. Level of MDM Maturity No vision. Firefighting is the answer. Isolated, bottom-up initiatives. Okay, let’s do something at the silo level. Silo-oriented solutions. No vision; But, “yes, we do have a problem” Problem? What Non-existent 1 Initial 2 Developing 3 Defined 4 Managed 5 Optimizing Maturity Stages Gartner
29
The MDM Process Life Cycle
Author Store Pub/ Synch Enrich Consume Archive Collaborate Enrich Enrich Enrich Enrich Enrich E2E Marketing Procurement Operations Logistics Sales Service Example Life Cycle for Product Master Data Questions to Answer: What processes will we need to ensure the creation, management, publishing and leveraging of high-quality master data across our organization? What business processes will the master data life cycle processes support? Gartner
30
The Basis of Achieving Buy-In and Creating the Metrics
Organizational Structures, Roles, and Responsibilities Executive-Level Sponsor Information Governance Board MDM Team Centralized or Distributed MDM Infrastructure Team Security Privacy System Mgmt. App. Dev. Integ. Modeling / Metadata Info. / Architect Data Quality Monitoring / Reporting Data Steward Business IT Questions to Answer: Who creates and consumes master data? What are their roles? Do we have data stewardship roles and is it seen as a business responsibility? What organizational structure do we need to manage master data? How will we manage the change that comes with new ways of working? Gartner
31
Data Integration, Enterprise Data Warehouse
Why Data Quality? :15 Data Validation – When and Why? :30 Data Profiling Concepts – Why and How? 1:00 Data Integration
32
Pop Quiz Putting all of your data in a single place (platform) mean your data is integrated. True False
33
Data Integration Framework
Dimensional modeling (DM) is the name of a logical design technique often used for data warehouses. Dimensional modeling consists of conformed dimensions and fact tables. A conformed dimension is defined and implemented one time, so that it means the same thing everywhere it's used. Fact tables that should be conformed include those that derive expenses, enrollment, courses, prices, and adds / drops. CONFORMED DIMENSIONS
34
Conformed Dimensions Methodology
Achieve Data Integration through Conformed Dimensions and Facts Conformed Dimension Management is based upon best practices of Master Data Management and Data Warehouse Bus Architecture by providing a framework for the warehouses to grow and integrate through a common set of conformed dimensions and defined conformed fact entities and metric definitions across the enterprise Also provides advanced update management and metadata management features ensuring timely content management and control over strategic hierarchies for mission critical business processes CDM is based on the Data Warehouse BUS Architecture described below A BUS is a common structure to which everything connects and from which everything derives power. By defining a standard bus interface for the data warehouse environment, separate data marts can be implemented and can be plugged together and usefully coexist if they adhere to the standard. DATA WAREHOUSE BUS ARCITECHURE Drops/Adds Credit Hours Student Counts Term College Program Campus Instructor Course Location
35
UF Data Management Reference Architecture
Transactional Layer Integration Layer Distribution Layer UF Standard User Enterprise Systems Enterprise Bus. Intelligence Click Commerce D I R E C T BI Platform Meta Data _ Dashboards UF Advanced User People Soft Reporting Scorecards College / 3rd Party E S B Canvas ETL ODS Enterprise Data Warehouse Developer / Advanced Tools DQ Visualization Research Computing Analytics UF Developer A P I Big Data SQL / Direct Analytics MDM Data Steward MDM HUB API MDM Work Flow Data Architecture
36
Dimensional Modeling Reference Architecture
Staging Conformed Dimensions Facts Person ID College Course Num Program Dept
37
Data Model Creation Process
1. Conceptual Model 2. Logical Model 3. Physical Model Indexes, Partitions, Optimization Student Student ID Name Type Status …. Student Student Student ID Name Type Status …. Schedule Date Course ID Term ID Student ID … Course Schedule Schedule Date Course ID Term ID Student ID … Term Catalog Governance Council Approval Data Steward Group Approval Architecture Group Approval
38
Business Process Flows
Student Lifecycles Retained Financial Aid Business Process Flows Advisement Recruitment Registration Degree-Seeking Student Admitted Alumni Bursar Admission Instruction Returns Next Term Returning Left (not retained) Registration Student Life Non-Degree Seeking Student Admission Bursar Finished Graduated Returns Later Instruction
39
Bursar Past Due Payments Student Disbursements Bursar Payments
Account Status From Past Due Payments Collects Full Name Student To Disbursements Bursar Distributes Payments Makes Permanent Address/Contact Info Program Indicators Tuition Status Account Status
40
Financial Aid Need/Eligibility Financial Aid Officer Student FAFSA
Need Based Eligibility determines Need/Eligibility Financial Aid Officer Student FAFSA Submits Entered/ Validated by Award Letter sends Application UFID Emergency Contact Full Name Maiden/Former Name SSN Gender Permanent Address/Contact Info Date of Birth Ethnic Origin Race Indicators UF Address/Contact Info Approved by Academic Department Permanent Address/Contact Info
41
RAPID ACCESS INSTANT Student View Student 360
Consolidate students across colleges, departments, degree programs, affiliates Student View ACCESS RAPID Identify students in the same household Identify students with multiple billing accounts or mode of payment Allow attribute-sharing across sources for consolidated students MDM:CDI Party Concept - Party Identifier allows studens, faculty, staff, etc. to be identified as one party having multiple roles Student 360 Holds Interests INSTANT Degree Programs Extracurricular
42
MDM Data Model – Party Concept (Student)
Consolidate students across campuses, programs, colleges, etc. Identify students in the same household Identify students with multiple billing accounts or mode of payment Allow attribute sharing across sources for consolidated student SSN might be in source 1 only but after consolidation shared with all sources Party Identifier allows students, faculty, staff, etc. (roles) to be identified as one party having multiple roles Students within the same household can be identified with one party id. Students with multiple billing accounts can be identified with one party id.
43
Student 360 Overview Diagram
Semester Flow Degree Program Course Registration Enroll Register Session Begins Attend Class Mid-Term Grade Final Grade Session Ends Drop Add Integrated Person Hub Course Schedule Course Credit Location Time Instructor Revenue Metrics Registration Fees Course Fees Credit Hours Student Fees Financial Aid Rich Data Enrollment Date Demographics Credit Hours Revenue Propensity to Churn Firmagraphics Scores Profitability Class Location Student Holds Campus Building Classroom Type of Hold
44
Data Integration, Enterprise Data Warehouse
Why Data Quality? :15 Why Data Quality?
45
Pop Quiz Does data quality matter more in your : Master Data
Data Warehouse Reporting Tools
46
ALL OF THE ABOVE ..and the Answer is….
Does data quality matter more in your : Master Data Data Warehouse Reporting Tools ALL OF THE ABOVE Data Quality Matters
47
One is its inherent quality, and the other is its pragmatic quality.
What is Data Quality? There are two significant definitions of information quality. One is its inherent quality, and the other is its pragmatic quality. Inherent information quality is the correctness or accuracy of data. Pragmatic information quality is the value that accurate data has in supporting the work of the enterprise. Data that does not help enable the enterprise to accomplish its mission has no quality, no matter how accurate it is. “Experience is revealing that more than half of data warehouses built fail to meet expectations because of poor information quality.” -Improving Data Warehouse and Business Information Quality,
48
Data Quality Framework
Process Quality Information Quality Systems Monitoring Performance Monitoring Mainframe Audit File Management ETL Audit Load Audit Data Mart Audit Reporting Audit Data Procurement Analysis Feed Validation Data Integration Methodology ETL Data Validation Data Certification Quality Reporting Conditional Reporting Source System Loop Back Multi-Source System Validation Work Flow Subject Matter Expertise Architecture & Integration Issue Tracking
49
Data Element Maturity Levels
Best of Breed Data Asset Better Data Asset Quality Data Asset Data Asset Understood Data Identified Data Confidence and Benchmarking Critical Data Elements Consistent Improvement and Usage Achieving Quality Expectations Policy and Rules Defined Owned, Modeled and Defined
50
Data Profiling 101 Basic Column-Level Analysis
Distinct count and percent Zero, blank, and NULL percent Minimum, maximum, and average string length Numerical and date range analysis Key integrity Cardinality (e.g. one-to-one, one-to-many, many-to-many, etc.) Pattern, frequency distributions, and domain analysis (e.g. Data Profiling Profiling normally examines areas such as data values, value ranges, frequency distributions, metadata mismatches, various statistics, non-standard record formats, etc
51
Data Profiling Worksheet
Column Name # of Records Inferred Data Type Distinct Count Null Count % Null Maximum Minimum # of Patterns Mean Median Standard Deviation Column Profiling Frequent values, outliers, maximum, minimum, nulls, patterns, overloaded use Table and Cross-Table Profiling Dependencies, candidate primary keys, candidate foreign keys, cardinality of relationships, referential integrity Additional Analysis Mean, median, standard deviations, uniqueness, ranges, reasonableness
52
Importance of Advanced Data Quality (Matching Algorithms)
Better data hygiene drives better data matching Better matching drives better student identification and modeling Better identification and modeling drives better student interactions Better interactions and campaigns drive higher retention rates Higher retention rates drive more revenues. Not doing MDM may be more expensive than doing it!
53
Sample Student Data - Today
*Fictitious Student Information Cust. Id First Name William Middle James Last Name Sosulski DOB April 12 Phone Address 123 Oak St., Eves, IL 30319 Recruiting Registrar Library Campus Security Financial Aid Cust. Id First Name William Middle J. Last Name Sosulski DOB Phone Address 123 Oak St., Eves, IL Cust. Id 14239 First Name Bubba Middle J. Last Name DOB April 12 USER # vz1234 Address Cust. Id 3721B First Name Willaim Last Name Corp Userid vz1234 DOB 04/12/1939 Account Address 3224 Pkwy G, Los Osos Cust. Id First Name William Middle James Last Name Sosulski DOB 04/12/1939 Account Address 123 Oak St., Eves, IL 30319
54
After Matching – Master Person View
Cust. Id First Name William Middle James Last Name Sosulski DOB April 12 Phone Address 123 Oak St., Eves, IL 30319 Cust. Id First Name William Middle J. Last Name Sosulski DOB Phone Address 123 Oak St., Eves, IL Cust. Id 14239 First Name Bubba Middle J. Last Name DOB April 12 USER # vz1234 Address Cust. Id 3721B First Name Willaim Last Name Corp Userid vz1234 DOB 04/12/1939 Account Address 3224 Pkwy G, Los Osos Cust. Id First Name William Middle James Last Name Sosulski DOB 04/12/1939 Account Address 123 Oak St., Eves, IL 30319 Person Hub 1001 William James Sosulski 04/12/1939 123 Oak Street Eves CA 91403 *Fictitious Student Information Recruiting Registrar Library Campus Security Financial Aid
55
Data Governance Best Practices
:30 Data Governance Data Governance
56
Data Governance Operating Model Overview
Data Trustees Data Trustees & Sponsors Typically core office(s); oversees all business function(s), DDD level, business-driven. Commonly responsible for data content, context, and associated business rules. Standards & Procedures Common communication and process mechanisms used to guide efforts and decisions Work Groups (by Focus) Data Governance Council Data Governance Council Establishes and manages governance team structures Potential issue escalation path Student Data Governance Council Data Custodians (IT) Data Stewards Data Custodians (adhoc) Typically IT-focused; can be institutional or local; responsible for the safe custody, transport, storage of the data and implementation of business rules. Data Stewards Typically core office; day-to-day functions; adding / editing relevant data, data quality assurance, specific business function(s).
57
Data Governance Council
The Data Governance Council is comprised of key stakeholders across the University who play an active role in the development and management of research and sponsored project information. Establishes overall policy and guidelines for the development of standards, definitions, classification and use of EM’s master data; Charters Working Groups to review and document definitions, hierarchies, taxonomies, data standards, business rules, sources of truth, and other meta-data associated with master data under the stewardship of EM. Coordinates EM’s interest in the definition, standardization and classification of enterprise-wide master. Monitors quality and accountability to standards. Student Data Governance and Master Data Management Charter, page 6
58
Data Ownership vs. Data Stewardship
Data stewards take care of the data Data stewards know the content Stewards are responsible for the quality Operate on behalf of the organization A Business function, not IT The farmers take care of the land Nobody owns the data or its use The owner is the organization It is an abstract concept The King/Queen owns the land
59
Data Stewardship Roles and Benefits
Data Steward generically refers to the four types of data stewardship committee roles: Executive Sponsor Any initiative that cuts across a company's lines-of-business must have executive management support onboard. Chief Steward Responsible for the day-to-day organization and management of the data stewardship committee. Business Steward Responsible for defining the procedures, policies, data meanings and requirements of the enterprise. Data Steward (Technical) A technical person that is a member of the organization's IT department. Benefits of Data Stewardship Consistent use of data management resource Easy mapping of data between computer systems and exchange documents It has been found that people are more likely to trust and use a system where there is a person they can call with question on each data element
60
Data Stewardship Responsibilities
A data steward ensures that each assigned data element: Has clear and unambiguous data element definition. Does not conflict with other data elements in the metadata registry (removes duplicates, overlap etc.) Has clear enumerated value definitions if it is of type Code. Is still being used (remove unused data elements) Is being used consistently in various computer systems Has adequate documentation on appropriate usage and notes Documents the origin and sources of authority on each metadata element Data Stewardship Framework by David D. Marco
61
Building the Winning Team
Work Groups The Work Groups will be accountable for defining standards and imparting data-centric knowledge, business representation, and data quality. Work Groups may consist of Data Stewards who will be responsible for communicating and imparting the decisions made on data domains data quality and data usage. The Work Groups will make recommendations on data changes and will bring the recommendations to the DGC for review, approval and execution. Building the Winning Team
Similar presentations
© 2025 SlidePlayer.com Inc.
All rights reserved.