dwdw dwwgrespons-20070529b-1
TRANSCRIPT
-
7/28/2019 dwdw DWWGRespons-20070529b-1
1/24
A Roadmap for
Data Warehouse,
Reporting and Analyticsat Georgetown
A Response to the Report of the Data
Warehouse Working GroupRon Allan, Dave Lambert, Matt McNally, Piet
Niederhausen
-
7/28/2019 dwdw DWWGRespons-20070529b-1
2/24
The growing importance of data in the life of theinstitution
Introduction to key concepts ..developing a common vocabulary
Some important issues facing forward
What can we learn from the experience of others
Addressing some short term tools issues
A proposed roadmap forward
A discussion about resources
Overview
-
7/28/2019 dwdw DWWGRespons-20070529b-1
3/24
Increasingly our executive leaders are demanding informationand analysis to support strategic decisions.
To assure efficiency and competitiveness of our day-to-day
operations, managers and directors require a constant flow ofreliable information.
Hardly a month passes that a new reporting requirement isntimposed from an external source.
..and even more are on the way.
Georgetown, like the rest of higher education, is being hit by aperfect storm of data issues.
But
Building the information-driven
university
-
7/28/2019 dwdw DWWGRespons-20070529b-1
4/24
Complexity of cataloging all data fields
Complexity of authorization: the Who sees what?
Complexity of access: the How do I see what?
No single management or reporting tool meets therange of information access needs
Regulatory issues: FERPA, HIPAA & G-L-B
Data security
Data accuracy issues: Sarbanes-Oxley
Confusing concepts and vocabulary
There are a multitude of
complicating issues
-
7/28/2019 dwdw DWWGRespons-20070529b-1
5/24
Common concepts and
vocabulary From the initiation of our first efforts in building data
warehouses, we have been hampered by a confusion inconcepts and vocabulary.
We are not alone in that regard There are still major disagreements among thought leaders (including
academics)
Vendors have unleashed an array of products with conflicting namesand capabilities.
One vendors warehouse is anothers ODS. There is every evidence that confusion will get muddier rather than
clearer with the accelerating consolidation in the market.
-
7/28/2019 dwdw DWWGRespons-20070529b-1
6/24
Reporting and analytics
Business systemsLook up and edit individual records.
Execute tasks and record transactions.
Operational data storesAccess data from one business domain.
Create production reports.
Integrated data environment (e.g., EDW)Correlate data across systems.
Report on historical data over time.
Create university-level reports.
AnalyticsCreate metrics based on institutional goals
Monitor performance using dashboards.
Conduct institutional research.
-
7/28/2019 dwdw DWWGRespons-20070529b-1
7/24
Business systems
Business systems contain transactional dataupdated by staff and end-users via self-service.
Access control is fine-grained (down torecords and fields).
Detailed knowledge of a system is required
to manage its data. How data is collected and stored affects the
ability to do reporting later.
-
7/28/2019 dwdw DWWGRespons-20070529b-1
8/24
Operational Data Stores
Each data store contains a copy at a point-in-time of transactional data from a
business domain, updated periodically.
Data stores are primarily used forproduction, line-of-business reporting.
Access control is modeled on business
systems, managed by database notreporting tools.
Domain knowledge and reporting skills arerequired to create reports.
-
7/28/2019 dwdw DWWGRespons-20070529b-1
9/24
Integrated data environment
Contains data extracted from many business domains, storedover time.
Some data is transformed to make key data consistent acrossdomains.
Multiple reporting tools should be supported.
The same access controls should be enforced across all
reporting tools. Access control is inherently less fine-grained than
underlying business systems to enable broader vision.
Reporting is constrained by how the data was originallycollected and stored.
Extraction,
transformation,
and load
(ETL)
Consistent
access
controls
Reporting
tools
Data stores Integrated data
environment
-
7/28/2019 dwdw DWWGRespons-20070529b-1
10/24
Analytics
Goals must be defined to establish measurableperformance metrics and research questions.
An analytics environment provides dashboardsthat show current status compared with a defined
goal. An analytics environment also provides the ability
to run hypothetical scenarios and projections forinstitutional research.
Institutional,
program, and
project goals
Performance
metrics
Research
questions
SVP Provost VP Dean
Dashboards
Analytics Scenarios Projections
-
7/28/2019 dwdw DWWGRespons-20070529b-1
11/24
Differentiators
Intended for Type of data Access to data
Required
expertise
Business
systems
Staff and
administrators
Individual
records and
transactions
Fine-grained
access controls
Detailed
knowledge of
base system
Operational
data stores
Staff and
administrators
Production data
from one
domain
Domain specific
fine-to-medium-
grained access
controls
Domain
knowledge and
reporting skills
Integrated
dataenvironment
Administrators
and executives
Data across
systems, overtime
Access by
domain orbusiness entity
Domain
knowledge
Analytics Executives and
institutional
research
Metrics and
trends over
time
Broad, high-
level access
High-level
domain
knowledge
-
7/28/2019 dwdw DWWGRespons-20070529b-1
12/24
Overview
Business systems
Operational data
stores
Integrated data
environment
Analytics
Student
Advance-
ment
Human
Resources
Endowment
Mgt
Financials
Research
Mgt
Benefits
Space/
Facilities
Faculty
Service
Mgt
Student
Advance-
ment
Human
Resources
Endow-
ment Mgt
Financials
Research
Mgt
Benefits
Space/
Facilities
Faculty
Service
Mgt
Reporting
tools
Institutional,
program, and
project goals
Performance
metrics
Research
questions
Dashboards
Analytics
Data access
policy
Data
administration
Data
governance
-
7/28/2019 dwdw DWWGRespons-20070529b-1
13/24
Status
Business systems
Operational Data
Stores
Integrated data
environment
Analytics
Student
Advance-
ment
Human
Resources
Endowment
Mgt
Financials
Research
Mgt
Benefits
Space/
Facilities
Faculty
Service
Mgt
Student
Advance-
ment
Human
Resources
Endow-
ment Mgt
Financials
Research
Mgt
Benefits
Space/
Facilities
Faculty
Service
Mgt
Reporting
tools
Institutional,
program, and
project goals
Performance
metrics
Research
questions
Dashboards
Analytics
Data access
policy
Data
administration
Data
governance
-
7/28/2019 dwdw DWWGRespons-20070529b-1
14/24
Issues
Establishing a culture and organizational focal point for information-based analytics in a consensus system
Performance analysis
Institutional research
Assuring current and future systems support DW/Analytics. The Financial system is a particular focus
In reference to the model, the max above is constrained by the minbelow
Establish mechanism(s) to address:
Data access policy
Data governance
Data administration
Building a long-term funding strategy
Some foundational elements are in the investment plan
Missing the dedicated resources per application area
In both UIS and Functional areas.
-
7/28/2019 dwdw DWWGRespons-20070529b-1
15/24
Experience of Peer Institutions
MIT
George Washington (GW)
University of Pennsylvania (Penn)
Yale
-
7/28/2019 dwdw DWWGRespons-20070529b-1
16/24
Sample of Peers
MIT Penn GWU Yale
Data
Governance
No
Comprehensive
Approach
No
Comprehensive
Approach
No
Comprehensive
Approach
No
Comprehensive
Approach
DataAdministration
Tradit. Aspects(catalog,
definition
Covered
DesignatedData
Administrator
No DesignatedData
Administrator
Conducted byBusiness
Service Data
Users
Data
Warehouse
Best developed
in Higher Ed
Kimball
Well Developed
Moving from
Inmon toKimball
Well Developed
Kimball
Sporadically
Developed
Adhoc
Analytics/
Dashboards
Dashboards
Extent in Non-
Executive
Community
Dashboard
Analyst
Assigned
Board has 1
Dashboard
Managers have
Dashboards
Executives Do
Not
-
7/28/2019 dwdw DWWGRespons-20070529b-1
17/24
Sample of Peers
MIT Penn GWU Yale
Data
Ownership
Not an Issue:
Trustees respn
for all aspects
of data
Institution Owns
the Data
Data Owned by
Relevant
Individual
Data Owned by
Relevant
Business
Manager
Data Access Granted byTrustee
Granted byData Steward
Granted byData Owner
Granted byData Owning
Departments
Staffing 7 FTEs in
Central
Technology Org
11 FTEs in
Central
Technology Org
5 FTEs 12 FTEs in
Central
Technology Org
Funding
Model
Formal Part of
IT Budget
Ad Hoc:
Overhead
funding to date
Combination of
formal and
project funding
Formal Budget
Process
-
7/28/2019 dwdw DWWGRespons-20070529b-1
18/24
Conclusions from peer analysis
Peer institution beset by the same demands andchallenges.
Many have launched data-related initiatives similar tothose at GU.
Most have made more progress on dataadministration and access policy.
Most have moved further up the integrated DWlayer..
..but with selected data
We could not find a really good exemplar of an R1
university that has a comprehensive solution
-
7/28/2019 dwdw DWWGRespons-20070529b-1
19/24
Status of our current toolset
ETL tool: Informatica
Top of the line
Query Tool: Cognos Version out of date
Web reports successful
Security is administered in Cognos rather than underlyingdatabases
This is a significant constraint
We have no standard tool for Analytics at this point
SAS is often used for statistical processing
Nothing available for true analytics
-
7/28/2019 dwdw DWWGRespons-20070529b-1
20/24
Open tools issues
ETL tool: Stay with Informatica or move to OraclesWarehouse Builder
This is Banners tool
Query Tool: Cognos
Given the market is in flux we can:
Stand pat for awhile (3 years?)
This is what SunGardHE decided to do
Would require upgrade to latest release
Tool choice for analytics
SAS is the industry leader, but there are others we shouldlook at:
Hyperion, Banner Analytics, etc.
-
7/28/2019 dwdw DWWGRespons-20070529b-1
21/24
Addressing tools issues
Establish Analytics (business intelligence)tools evaluation committee.
Make recommendations on ODS and IDE tools:Investigate Analytics tools options.
Lay the ground work for next generation querytool choice
As market stabilizes.
Members from UIS and user communities.
-
7/28/2019 dwdw DWWGRespons-20070529b-1
22/24
Proposed Calendar (1)
Replace SIS with Banner
Define FMS then Replace
Stabilize & Enhance Space
Migrate HRIS off Mainframe
2007 2008 2009 2010 2011Business Systems Context
Review interim ETL and reporting tools
Establish Secure Reporting Environment
Deploy 2nd Generation ODSs for
Student, Financials, HR, and Space
2007 2008 2009 2010 2011Refresh or Build ODSs
-
7/28/2019 dwdw DWWGRespons-20070529b-1
23/24
Proposed Calendar (2)
Establish Data Administrator
Build and Hire Data Admin team
Develop 1st generation data inventory
Establish Data Governance committees
Develop Data Access policies
2007 2008 2009 2010Data Administration
Develop 1st generation IDEs
Develop 1st generation Analytics
Integrated data environment & Analytics
Hire IA and Reporting Vacancies
UIS and Business Division Staffing
-
7/28/2019 dwdw DWWGRespons-20070529b-1
24/24
Organizational Model
Data Access, Analytics, Self Service, Enterprise Web
Associate Director,
Information Access
Web & Data
Architect
BusinessAnalyst
Data
Engineer
Data
Engineer
Application
Programmer
DataAdministrator
Application
Programmer
Technical Manager,
Information Access
Application
Programmer
Director, EETS
DBA
AVP, NCS
DBAsApplications, Managers
Financials, HR, Space, Student
Application
Teams
Data Access
Policy
Working
Group
Data
Governance
Committee
Business
Reporting
Analyst
Business
Reporting
Analyst
Business
Reporting
Analyst
Business
Reporting
Analyst
Business
Reporting
Analyst
Business
Reporting
Analyst
Business Divisions
Business Divisions