Successful Data Warehouse Implementation

advertisement
Successful Data Warehouse Implementation
Brings Changes to Campus Culture, Processes,
and People
Ora Fish
Rensselaer Polytechnic Institute
Agenda
Alignment between the Information
Technology and the Business
Development Methodology
Lessons Learned
Demonstration
Q & A
Rensselaer
Polytechnic Institute
(RPI)
Founded in 1824 by Stephen Van Rensselaer
• “We are the first degree granting technological
university in the English-speaking world”
• Research University
• Total Students 8,265
– Graduates: 1,378
– Undergraduates: 5,164
– Education for Working Professionals: 1,723
• Faculty - 450
Data Warehouse group
• Part of the Administrative Computing
within the Division of Chief Information
Office
• Total of eight employees
• Responsible for addressing campus
reporting and analytical needs
• http://www.rpi.edu/datawarehouse/
Fundamental Problem
Operational systems are not designed for
information retrieval and analytical
processing
The Fundamental Goal
The fundamental goal of the Rensselaer
Data Warehouse Initiative is to integrate
administrative data into a consistent
information resource that supports
planning, forecasting, and decisionmaking processes at Rensselaer.
Data Warehouse Objectives
• Serve as an information hub for Administration
as well as the Academic Schools
• Transform Data into Information with embedded
business definitions
• Informative - Meta Data
• Intuitive for end user to perform ad-hoc queries
and analysis
• Adequate response time - Retrieved within
seconds
Implementing Data Warehouse
Alignment
Business
Technology
Information
Quality
Campus
Culture
Information Quality
Accurate, Reliable, Consistent, Relevant
• Re-enforce common definitions
• Set up processes to identify and clean
erroneous data
• Set up processes to gather relevant data
• Define policies on who will have access to
what information
Culture
• Promotes fact based decisions
• Requires lowering the walls across
organizational boundaries
• Understanding the business enterprise
across different functional areas
• Analytical culture requires different set
of skills
Before, During, and After the
implementation
How does the IT leads and effectively
aligns Technology, Information Quality,
and Campus Culture before, during,
and after the Data Warehouse
implementation
Implementation Methodology
Campus Communication
Build DW
Foundation
Develop
Subject Oriented
Data Marts
Release Data Mart
To the Core Administration
Data stewards
Training
Release Data Mart
to the Campus
Adaptation and
Growth
Maintenance and Support
DW Program Timeline
FY02
Infrastructure Planning/Staffing
Software
Database/Hardware
Production Platform
Policy
Data Policy
Datamarts
Finance/Research
Position Cntrl/Labor
Human Resources
Enrollment
Grad Financial Aid
Undergrad Fin Aid
Admissions
Contracts & Grants
UG Prospect Fin Aid
PI/Fund Info Access
Alumni/Advancemnt
Operations Support
Software Upgrade
Database Upgrade
Hardware Growth
Req
FY03
Dev & Test
FY04
Rollout
FY05
Building DW Foundation
• Organizational Structure
• Project framework and high level
plan
• Building Technical Infrastructure
• Develop Data Policies and
Procedures
Project Organizational
Structure
Alignment between the IT and the Business
Sponsorship Group
(Business & IT)
Approves Next Areas
Steering Committee
(Business & IT)
Forming Implementation groups; Defining scope and deliverables
Implementation groups
(Business & IT)
Implementation issues
Data Warehouse Group
(IT)
High Level Analysis and
Prioritization process
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Enro
ll me
nt M
gm t
Stud
e nt L
ife
Ins ti
tute
Ad v.
Prov
o st O
ffice
Rese
arch
Fin a
nce
Gov
' t Re
la tio
ns
Pres
i den
ts St
aff
HR
BUSINESS PROCESSES
Enrollment Analysis
Student Pipeline Analysis
Faculty Workload Assessment
Financial Analysis
Contract and Grants Analysis
Proposal Pipeline Analysis
Financial Analysis - Research
Graduate Financial Aid
Alumni Demographics and Tracking
Alumni Contact Management
Human Resources
Facilities Management
Sch o
ols/ D
e ans
Dep'
t Ch
ai rs
Regi
strar
CONSTITUENCIES
X
X
X
X
X
X
Prioritization Process
High
PP AC
FR
FA
SP
CG
HR
Value to
Rensselaer
(Culture)
Low
FW EA
AD
GF
Feasibility (determined by
Data Quality and availability)
FM
High
Technical Architecture Inventory
•
•
•
•
ERP – Banner from SCT
ETL – Power Center from Informatica
Data Base – Oracle 9i
Models – Star schemas with conformed
dimensions
• Web Front end tools – Brio, Dash Boards
• Desktop Front End tools – Brio, Excel
Data Security, Privacy and Access Policy
Security
&
Privacy
•
•
•
Access
& Use
Can be defined as striking the “right” balance between
data security/privacy and data access
Value of data is increased through widespread access
and appropriate use, however, value is severely
compromised by misinterpretation, misuse, or abuse
Key oversight principle:
–
Cabinet members, as individuals, are responsible for overseeing
establishment of data management policies, procedures, and
accountability for data governed within their portfolio(s), subject
to cabinet review and CIO approval
Building Subject Oriented Data
Marts
•
•
•
•
•
•
•
•
Alignment between the Technology, Information Quality, and
Campus Culture
Determining Constituency • Identify information gaps
Forming Implementation • Identify erroneous data
Group
• Reinforce common definitions
Conducting interviews
• Establish processes to identify and
clean erroneous data
Defining Scope and
• Establish processes to capture
Timelines
missing data
Modeling
• Develop and approve Data Security
Extracting, Transforming,
Policy
and Loading Data
• Record Meta Data – stored in
Develop Security system
Informatica repository and accessed
with Brio
Testing
Modeling: Kimball’s Bus Architecture to Subject oriented Data
Marts and the Conformed Dimensions
Student Enrollment Model – one row per enrolled student per term
Student Enrollment Snapshot
Academic Term
Dimension
Academic Term Key
Student Dimension
Student Key
Student Enrollment
Snapshot
Academic Program
Bridge Dimension
Student Academic
Program Key
Class Dimension
Academic Term Key
Prim Program Major Grp Key
Sec Program major Grp Key
Student key
Student Cohort Key
Class Key
Student Faculty Advisor Key
Class Key
Student count
Matriculated count
Credit Hours Registered
Credit Hours Attempted
Credit Hours Earned
Overall GPA
Term GPA
Tuition Amt Charged
Tuition Fees Charged
Tuition Amount Billed
Tuition Fees Billed
Student Cohort Key
Student Cohort
Bridge
Student Faculty
Advisor Bridge
Student Faculty Advisor
Key
Faculty Advisor Key
Cohort Key
Highlights of the Development –
ETL Process
In addition to the data staging and development
processes:
• Develop Data Quality Assurance Processes
• Ensure transformations are captured
• Capture data at the lowest level – no one ‘trusts’
statistics only without the supportive details
• Initial and Incremental Loads
Development - Securing Data
Marts
• Working with each portfolio, the IT role was to
ensure that the subject oriented Data Policy is:
Defined, Approved, Technically feasible, and
Consistent Across The Board
• Build Security Front End application
• Security Options:
–
–
–
–
Securing schemas
Securing facts only
Securing dimensions only
Securing both facts and dimensions
Securing Facts Only
Time Dim
Time Key
Calendar Year
Calendar month
Calendar day
Date
Fiscal year
Fiscal Period
Account Dim
Acct Key
Acct code
Acct Description
Acct Type
Various Indicators
Various attributes
Financial Transaction Fact
Org Key
Fund Key
Acct Key
Time Key
------------------------
$
Organization Dim
Org Key
Org code
Org Description
Org Financial Manager
Type
Various Indicators
Various attributes
Fund Dim
Fund Key
Fund code
Fund Description
Fund Financial Manager
Type
Various Indicators
Various attributes
Securing Only Identifiable
Information
Time Dim
Time Key
Calendar Year
Calendar month
Calendar day
Date
Academic Term
Academic Year
Snapshot Type
Student Dim
Std Key
Name
Demographics
Geographic Info
Minority Ind
Citizenship reporting Ind
Various Indicators
Various attributes
Student Advisor Bridge
Advisor Key
Name
Weight Factor
Type
Rank
Student Enrollment Fact
Discipline attributes
-----------------------Title
Count
School
Matriculated Count
Department
Various attributes
Hours Registered
Academic Program Dim
Hours Attempted
Academic Prg Key
Credit Hours Earned
Type
Term GPA
Classification
Overall GPA
Major/Minor/Concentration Type
Tuition Paid
Major/Minor/Concentration Desc.
Campus
Etc….
School
Department
Official Headcount Indicator
Weight Factor
Various attributes
Class Dim
Class Key
Class code
Class Descr
Managing Testing Sessions
Alignment between the Technology, Information Quality, and
Campus Culture
• Allocating time slots
• Focused – aiming to produce existing reports
and Queries
• Verifying that the models do address the need
• Opportunity to create more definitions,
groupings, and transformations
• Great opportunity to bridge diverse groups
• Further Enforce Common Definitions
• Further Identify Information Gaps
Campus Communications
During Testing Period
Meeting one-on-one with Campus
executives (Cabinet, Deans, etc.)
– Getting feedback early on
– Engaging
– Marketing
Data Mart Release to the Core
Administration/Data Steward
• Utilizing Data Mart for internal operations
• More changes to the Data Mart are expected
Information
Quality
• Establishing data cleanups queries and procedures
• Preparing for Campus release:
Impacting
Culture
– Developing campus training program: Developing and publishing
Dash Boards, and Brio dynamic documents
– Developing operational training
Campus Rollout
Alignment between the Technology, Information Quality, and
Campus Culture
•
•
•
•
Developing Roll-Out strategy
Defining roles and responsibilities
Defining initial access level
Recognize barriers and Setting
expectations
• Designing Training Programs
• Communicate to the Campus
Data Warehouse Cascaded
Rollout Strategy
1. Core Administration
2. Portfolio Level
(Cabinet,
Deans, Portfolio Managers)
3. Department Level
(Directors,
Center Directors, Department Chairs,
Department Financial Managers)
4. Other
Initial Tiered Access – Who
will have access to what
Cabinet; Deans;
Department Chairs;
Center Directors
Dash
Board
Department level
Information
published
In Brio
documents
Core Administration
Portfolio/Division level
Data in the Warehouse
Recognizing Barriers
• People’s resistance to a new tool
• Expectations on information availability
and usability for decision making are low
• Habit of relying on Central Administration
to provide information, or on their own
sources (many versions of the ‘truth’)
• People will need to acquire new job skills
• Job expectations will need to change
Developing Common Vision
• One version of the truth – Warehoused
Information was recognized as the only
official source of data
• Data Experts across campus and across
organizational boundaries
• Partnering with Human Resources – The
DW training was included in Performance
Evaluations and Job Descriptions
• Training is mandatory at all levels
Communication
• Executive briefings:
– Emphasized changes in analytical culture
– Recognized Barriers
– Emphasized that top down approach is needed and ask for
commitment
– Demonstrated new capabilities via Dash Boards
– Demonstrated ad-hoc capabilities people within their
organization have
• Campus orientations
– Demonstration were carried out by the original testing group
– Introduced training programs and the rollout strategy
– Communicated Data Policies
• Wed site
Training Mix
• Brio 101
– Basic navigation and
mechanics
• Brio 201
– Advanced analytics and reports
• Data Training
– Data mart basics, BQYs, and
star schemas
• Operational Training
– Focuses on practical
applications , delivered by
business owners
• Study Halls
– Informal, open agenda
• Best Practices
– Demonstration of best
practices, delivered by
business owners
• One-on-Ones
– Used to address specific
reporting/analytical needs
Training Program Overview
Track 1
Brio 101
Level 1:
Data Mart
Basics
Level 2:
Advanced Brio
Documents
High
Operational
Training
Track 2
Medium
Track 3
Low
Brio 101
Level 1:
Portfolio/Dept-Specific PreBuilt Docs
Dashboard & Portal training
One-on-one or small group format
Ongoing Follow-up
Training Philosophy
• The goal of the training program goes
beyond teaching the mechanics:
– Need to sell the Brio tool and the project
– Need to educate on the benefits of the DW
– Need to emphasize that Banner and the
DW are complementary systems, i.e.,
• Need to continue and inspire!
Adaptation and Growth
Adaptation and Growth
The true benefits can be achieved only
when the new technology is adapted
and becomes part of our business
routine:
• Penetration takes time
• Brings transformational changes to:
Processes and culture
Adaptation and Growth
Changes in our Processes
Some examples on utilization of the warehoused information in
our operations:
Assessment and Planning
• Enrollment Planning Committee meeting utilizes the
enrollment and the admission data in setting the enrollment
targets and financial aid goals as they discuss the incoming
class (how we did, quality, numbers, diversity, etc)
• Retention analysis – analyzing the admissions data to better
understand how well the incoming class may be retained next
year
• Assessment of Employee retention
• Assessment of Faculty renewal program
Adaptation and Growth
Changes in our Processes
Forecasting:
• Forecast current year sponsor research expenditures.
• Forecast graduate financial aid commitments
• Utilize past enrollment, retention, and financial aid information to forecast current and
future year financial aid commitments to determine the affordability of various discount
rates
• More accurately forecast research awards
• Utilizing historical research ‘success rates’ in projecting cost sharing commitments
Monitoring and compliance:
• Daily monitoring of budgets and expenditures from higher levels down to the specifics
• Monitor and review project to date budgets
• Monitoring positions budgets vs. actuals and in conjunction with estimated future
earnings are accurately projecting balances
• Monitoring the allocation of graduate financial aid
Operations
• Financial information is used in preparing and analyzing the financial statements,
reconciling between the sub-ledger and general ledger, reviewing payroll allocations
• Credit card reconciliation
Adaptation and Growth
Cultural Changes
• Empowers decision-makers: getting
accustomed to information availability
• Promotes the “no walls” culture
• From ‘MY Data’ to ‘Our Information’ - Data
Stewards role in improving data quality,
integrity, and conformity
• Fact based decision making
• Redirecting costly personnel hours
• Enhancing institutional effectiveness
Lessons Learned
• Building Data Warehouse is far more than a
technical endeavor: the alignment between the
right technology, information quality, and
campus culture has to be addressed before,
during, and after the data warehouse
implementation through planning,
development, testing, rollout, training, and
adaptation stages
• Business Sponsorship – is a must
Lessons Learned
• Properly designed Organizational Structure helps to
navigate political obstacles
• Partnership with the Business users – build it alone and
they will never come
• Identify your business ‘Stars’ as early as possible
• JAD and RAD approaches are best fitted for the iterative
Data Warehouse development
• Dash Boards – unless it is visible it is not there
Questions ???
Ora Fish
fisho2@rpi.edu
?
Download