snakepit or shangri-la?

28
Snakepit or Shangri-La? Snakepit or Shangri-La? Kathleen Moore, Jessica Foster, Nancy Speck, Kathleen Moore, Jessica Foster, Nancy Speck, Carl Dickinson Carl Dickinson University of Rochester University of Rochester AIR Forum 2007 - Kansas City, MO AIR Forum 2007 - Kansas City, MO Issues and Potential Pitfalls Issues and Potential Pitfalls in Implementing a Student Data in Implementing a Student Data Warehouse Warehouse

Upload: davin

Post on 06-Feb-2016

52 views

Category:

Documents


0 download

DESCRIPTION

Snakepit or Shangri-La?. Issues and Potential Pitfalls in Implementing a Student Data Warehouse. Kathleen Moore, Jessica Foster, Nancy Speck, Carl Dickinson University of Rochester AIR Forum 2007 - Kansas City, MO. Presenters: Kathleen Moore, Assistant Provost - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Snakepit or Shangri-La?

Snakepit or Shangri-La? Snakepit or Shangri-La?

Kathleen Moore, Jessica Foster, Nancy Speck, Carl Kathleen Moore, Jessica Foster, Nancy Speck, Carl DickinsonDickinson

University of RochesterUniversity of Rochester

AIR Forum 2007 - Kansas City, MOAIR Forum 2007 - Kansas City, MO

Issues and Potential Pitfalls in Issues and Potential Pitfalls in Implementing a Student Data Implementing a Student Data

WarehouseWarehouse

Page 2: Snakepit or Shangri-La?

22

Presenters:Presenters: Kathleen Moore, Assistant ProvostKathleen Moore, Assistant Provost Jessica Foster, Assistant Provost for Institutional Jessica Foster, Assistant Provost for Institutional

ResearchResearch

Co-authors:Co-authors: Nancy Speck, Assistant Dean for Institutional Nancy Speck, Assistant Dean for Institutional

Research and Registrar Research and Registrar Carl Dickinson, Associate RegistrarCarl Dickinson, Associate Registrar

Page 3: Snakepit or Shangri-La?

33

About the University of RochesterAbout the University of Rochester Doctoral Research-Extensive institution located in Doctoral Research-Extensive institution located in

upstate NYupstate NY 6 schools: Arts, Sciences, & Engineering; Music; 6 schools: Arts, Sciences, & Engineering; Music;

Business; Education; Medicine & Dentistry; and Business; Education; Medicine & Dentistry; and NursingNursing

Fall 2006 Undergrads: 4904, Graduates: 3539, Fall 2006 Undergrads: 4904, Graduates: 3539, MD’s: 403 MD’s: 403

Decentralized philosophy Decentralized philosophy Some centralized administration existsSome centralized administration exists

Page 4: Snakepit or Shangri-La?

44

Why Was a Data Warehouse Why Was a Data Warehouse Needed?Needed?

Current student data system (ISIS) – homegrown Current student data system (ISIS) – homegrown “legacy“ system, 20 yrs old, no plans to replace for 3-5 “legacy“ system, 20 yrs old, no plans to replace for 3-5 yrs yrs

Process for retrieving data from the ISIS system:Process for retrieving data from the ISIS system: Natural programs are run by the Registrar’s office to create Natural programs are run by the Registrar’s office to create

SAS extracts SAS extracts IR then writes mainframe SAS programs to retrieve data from IR then writes mainframe SAS programs to retrieve data from

extract filesextract files . . . FOR ANY REPORT!!!! . . . FOR ANY REPORT!!!!

Time to get data for simple questions takes hours; Time to get data for simple questions takes hours; more complex projects can take weeksmore complex projects can take weeks

The idea of a data warehouse seemed like a dream The idea of a data warehouse seemed like a dream come true!come true!

Page 5: Snakepit or Shangri-La?

55

Getting Buy-In and Finding Getting Buy-In and Finding FundingFunding

CIO and University Registrar were instrumental in CIO and University Registrar were instrumental in getting the pilot underwaygetting the pilot underway

To implement the student DW project and other To implement the student DW project and other projects, a Director of Administrative Computing was projects, a Director of Administrative Computing was hiredhired

In initial stages, funding the development of the DW In initial stages, funding the development of the DW was a serious issuewas a serious issue

Central IT ultimately provided all funding for the pilot Central IT ultimately provided all funding for the pilot (over $200K to date)(over $200K to date)

Funding for continued development is under discussionFunding for continued development is under discussion

Page 6: Snakepit or Shangri-La?

66

Choosing a SystemChoosing a System Decision-drivers:Decision-drivers:

Web-based user interface Web-based user interface Sufficient flexibility for a complex institution Sufficient flexibility for a complex institution Related systems already in use or plannedRelated systems already in use or planned CostCost Ease of use Ease of use

Final choice: COGNOS 8 (with Oracle Final choice: COGNOS 8 (with Oracle database)database)

Page 7: Snakepit or Shangri-La?

77

Convening the Right PeopleConvening the Right People Steering Committee consisted of a core group Steering Committee consisted of a core group

of IT staff, registrars, and IR staffof IT staff, registrars, and IR staff Total of 10 to 12 membersTotal of 10 to 12 members

Page 8: Snakepit or Shangri-La?

88

Arcane Architecture, Confusing Arcane Architecture, Confusing AcronymsAcronyms

Developing the warehouseDeveloping the warehouse Different ways of thinking: IT vs. End-UserDifferent ways of thinking: IT vs. End-User Limitations of current system posed problems with Limitations of current system posed problems with

fitting some data into a DW hierarchyfitting some data into a DW hierarchy In the end, no real data hierarchy was adoptedIn the end, no real data hierarchy was adopted

Nomenclature of warehouse elements was confusingNomenclature of warehouse elements was confusing Variable names not always the same as in ISISVariable names not always the same as in ISIS Organization of the warehouse into Models, Organization of the warehouse into Models,

Dimensions, Facts, and Attributes is very differentDimensions, Facts, and Attributes is very different Many lengthy meetings! Many lengthy meetings!

Have to be willing to invest timeHave to be willing to invest time Meeting weekly at a set time made it possible to Meeting weekly at a set time made it possible to

meet project deadlinesmeet project deadlines

Page 9: Snakepit or Shangri-La?

99

Guerilla DatabasesGuerilla Databases Schools have created their own databases outside Schools have created their own databases outside

of ISIS of ISIS Transferring information from these auxiliary Transferring information from these auxiliary

databases to the warehouse is an on-going issuedatabases to the warehouse is an on-going issue Office of Special Programs Office of Special Programs Study AbroadStudy Abroad CIP and NY State codes  CIP and NY State codes  

Interest in the data warehouse is promoting use of Interest in the data warehouse is promoting use of one central reporting systemone central reporting system

Page 10: Snakepit or Shangri-La?

1010

Testing, Testing…Testing, Testing… Testing environment confusing! Testing environment confusing! 

Different web addresses Different web addresses Changes in the COGNOS tool due to upgradesChanges in the COGNOS tool due to upgrades Moving reports from one environment to anotherMoving reports from one environment to another

Finding time to test was a big issue (still is!) Finding time to test was a big issue (still is!) Core user group tried to do this on their ownCore user group tried to do this on their own Weekly group testing meetings in a computer lab Weekly group testing meetings in a computer lab

with IT staff available with IT staff available Commitment to making this a success was our Commitment to making this a success was our

motivatormotivator

Page 11: Snakepit or Shangri-La?

1111

… … and More Testingand More Testing Slow response time during group testing was Slow response time during group testing was

frustrating (the dreaded “spinning box/hour glass”)frustrating (the dreaded “spinning box/hour glass”) Verification process was hampered because data was Verification process was hampered because data was

initially refreshed only once a weekinitially refreshed only once a week Testing enlightened users and programmers to ISIS Testing enlightened users and programmers to ISIS

limitations, processing quirks, and the need to make limitations, processing quirks, and the need to make changes to daily business practiceschanges to daily business practices Going through this process will eventually make it easier to Going through this process will eventually make it easier to

transition to a new student systemtransition to a new student system Originally only tested data from 5 most recent terms Originally only tested data from 5 most recent terms

to make sure everything worked – eventually loaded to make sure everything worked – eventually loaded ~20 years worth of data~20 years worth of data

Page 12: Snakepit or Shangri-La?

1212

Data CleanupData Cleanup Testing revealed a number of problems with how Testing revealed a number of problems with how

ISIS data was storedISIS data was stored Student addressesStudent addresses Degree conferral and award datesDegree conferral and award dates

Data entry cleanup – making sure data were Data entry cleanup – making sure data were initially entered and coded consistently in ISISinitially entered and coded consistently in ISIS Take Five studentsTake Five students McNair studentsMcNair students

Page 13: Snakepit or Shangri-La?

1313

Changing Business PracticesChanging Business Practices Data from multiple areas requires good Data from multiple areas requires good

communication and strong inter-departmental communication and strong inter-departmental relationships relationships Ex., IPEDS ethnic Ex., IPEDS ethnic

Still in progress – items not tracked in ISISStill in progress – items not tracked in ISIS Remedial and developmental coursesRemedial and developmental courses Students in combined degree programs (3/2, Students in combined degree programs (3/2,

MD/PhD, etc.)MD/PhD, etc.) Each school defines a “registered student” differently Each school defines a “registered student” differently

and uses variables differentlyand uses variables differently

Page 14: Snakepit or Shangri-La?

1414

Training and SupportTraining and Support An on-going issue An on-going issue

Initially IT consultants trained core group; IT & core Initially IT consultants trained core group; IT & core group developed training documents for new users group developed training documents for new users    

Currently more demand for training than we have Currently more demand for training than we have time available to devote to it, which may be slowing time available to devote to it, which may be slowing adoption of the DW by new users adoption of the DW by new users

Two types of training might be appropriate: IT Two types of training might be appropriate: IT training on the tool and core group training on the training on the tool and core group training on the datadata

Identifying who will be responsible for training prior to Identifying who will be responsible for training prior to implementation is implementation is crucial crucial (this was a shortcoming in (this was a shortcoming in our project)our project)

Page 15: Snakepit or Shangri-La?

1515

Training We Currently OfferTraining We Currently Offer Two members of the core group offer brief presentations Two members of the core group offer brief presentations

followed by a hands-on session ~ 3 times per year followed by a hands-on session ~ 3 times per year Monthly hands-on work sessions in computer lab where Monthly hands-on work sessions in computer lab where

volunteers from core group help new users with queries volunteers from core group help new users with queries Created a listserv for core group and current users to post Created a listserv for core group and current users to post

questions, concerns, etc. questions, concerns, etc.  A “decision tree” of what model to use is given to all usersA “decision tree” of what model to use is given to all users

Models can consist of either unduplicated or duplicated Models can consist of either unduplicated or duplicated records per studentrecords per student

Users need to fully define the question being asked in Users need to fully define the question being asked in order to pick the correct model order to pick the correct model

Page 16: Snakepit or Shangri-La?

1616

Do you want a report of all

students who have a particular major?

(regardless of major1,2,)

Student Counts/Lists by major

What kind of report do you want?

Student Counts/Lists

Statistical data (e.g. Hours

earned, GPA) from end of term or cumulative?

Course registration information

(e.g.add, drop, grades)

Degree information

Student Degree

Current Student Hold information

Student Hold

Do you want a report of all

students who have a particular program?

(regardless of program1,2,)

Student Counts/Lists by program

Is it a University level (vs. college) report where it is important that the

student not be counted twice if

they have multiple programs?

Student Counts/Lists unduplicated

Student Data WarehouseWhich model do I choose?

Just want to look at the values in

the lookup tables.

Student Table Lookups

Registration changes within a term are required.

(1 or more actions for a

term)

Student Registration

Activity

Only most recent

registration information for

the term. Multiple records

for duplicate programs.

Student Course Registration

Program Activity changes

(e.g.major, minor, time status)

within a term.

Student Program Activity

07/20/2006

Each student program displayed

separately (use this if creating a

report by program)

Student Term Statistics

One student program line,

regardless of how many programs

Student Term Statistics

Unduplicated

Only most recent

registration information for the term. (Use

this one most of the time)

Student Course Registration Unduplicated

Framework model selection.vsd

Which model do I choose?

Page 17: Snakepit or Shangri-La?

1717

A New Way of ThinkingA New Way of Thinking New users initially experience a lot of frustrationNew users initially experience a lot of frustration

Different naming conventions for DW fields from Different naming conventions for DW fields from names they’re familiar with in ISISnames they’re familiar with in ISIS

The way data fields are grouped into models The way data fields are grouped into models involves an entirely new way of thinkinginvolves an entirely new way of thinking

DW users aren’t always ISIS users, so ISIS DW users aren’t always ISIS users, so ISIS documentation is also available in the DW for documentation is also available in the DW for structural reference and logic structural reference and logic

New DW users need to break out of their “this is New DW users need to break out of their “this is how we’ve always done it” mindset how we’ve always done it” mindset

Page 18: Snakepit or Shangri-La?

1818

““Hey, this thing is cool – and it Hey, this thing is cool – and it really works!”really works!”

Ad hoc reporting is MUCH easierAd hoc reporting is MUCH easier Makes longitudinal analysis much more manageable – Makes longitudinal analysis much more manageable –

can easily report across years without merging extract can easily report across years without merging extract files files

Cool projects completed using the DW:Cool projects completed using the DW: National Research Council study on PhD degree completionNational Research Council study on PhD degree completion Tuition transfer study Tuition transfer study Credit hour analysis Credit hour analysis Classroom scheduling Classroom scheduling Music school instrument groupingsMusic school instrument groupings

Currently looking into using DW to create our Fact Book Currently looking into using DW to create our Fact Book

Page 19: Snakepit or Shangri-La?

1919

SamplSample e

ReportReportss

Page 20: Snakepit or Shangri-La?

2020

Page 21: Snakepit or Shangri-La?

2121

Reality vs. Wish List: Reality vs. Wish List: RealityReality Initially our DW was a PILOT project – we now need to Initially our DW was a PILOT project – we now need to

secure more funding, IT resources, etc. to continue secure more funding, IT resources, etc. to continue developmentdevelopment

Finding some areas need a different structure, more Finding some areas need a different structure, more data fields, historical data, or custom/calculated fields data fields, historical data, or custom/calculated fields

Our IT development team is now committed to other Our IT development team is now committed to other on-going projects in addition to the Student DW on-going projects in addition to the Student DW

Considering hiring an administrator fully devoted to Considering hiring an administrator fully devoted to the DW to help with ongoing upgrades, the DW to help with ongoing upgrades, enhancements, and trainingenhancements, and training

Page 22: Snakepit or Shangri-La?

2222

Reality vs. Wish List: Reality vs. Wish List: Wish ListWish List The core team has created a “wish list” prioritized by The core team has created a “wish list” prioritized by

level of importance (currently has about 20 items): level of importance (currently has about 20 items): Things we wish we’d thought of earlier: Things we wish we’d thought of earlier:

Ex., General Subject Area field (Humanities, Social Ex., General Subject Area field (Humanities, Social Sciences, etc.) Sciences, etc.)

Who is a “registered student?”Who is a “registered student?” Things we want for the “second round” of the project: Things we want for the “second round” of the project:

Ex., Financial aid dimension Ex., Financial aid dimension Things that just need to be changed or fixed: Things that just need to be changed or fixed:

Ex., Grade value should be a fact and not an Ex., Grade value should be a fact and not an attribute attribute

Usage of SSN – who can access this?  Can it be included Usage of SSN – who can access this?  Can it be included in DW but “hidden” from most users? in DW but “hidden” from most users?

Page 23: Snakepit or Shangri-La?

2323

Our “Wish Our “Wish List”List”

Page 24: Snakepit or Shangri-La?

2424

In Conclusion: Takeaway In Conclusion: Takeaway PointsPoints

Things we think we did Things we think we did well:well:

Selected the right core Selected the right core group group

Chose a software Chose a software platform that fits our platform that fits our university’s needsuniversity’s needs

Core user group Core user group worked well with ITworked well with IT

Hired an independent Hired an independent consultant who was consultant who was familiar with DW familiar with DW structure and COGNOS structure and COGNOS tool to review our tool to review our business practicesbusiness practices

Where we missed the boat:Where we missed the boat: One key person added too late One key person added too late

to the project – missed to the project – missed perspectiveperspective

More explicit communication More explicit communication between IT and core users – between IT and core users – user group assumed IT fully user group assumed IT fully understood our practices and understood our practices and needs at a very detailed levelneeds at a very detailed level

IT needed to educate core users IT needed to educate core users more on the role of models in more on the role of models in the DW structurethe DW structure

Planning in advance for future Planning in advance for future restructuring of some modelsrestructuring of some models

Needed better planning for user Needed better planning for user training upfronttraining upfront

Page 25: Snakepit or Shangri-La?

2525

So… Snakepit or Shangri-La?So… Snakepit or Shangri-La?

Page 26: Snakepit or Shangri-La?

2626

85-90% of Rochester’s schools have now 85-90% of Rochester’s schools have now adopted the DWadopted the DW The more we demonstrate it, the more people get The more we demonstrate it, the more people get

interestedinterested Has increased communication among offices which Has increased communication among offices which

typically don’t work togethertypically don’t work together Forced us to review data policies and business Forced us to review data policies and business

practicespractices Has served as important preparation for adopting a Has served as important preparation for adopting a

new central student information system in a few yearsnew central student information system in a few years Conclusion?Conclusion?

Page 27: Snakepit or Shangri-La?

2727

It’s still a work in progress, but so far the It’s still a work in progress, but so far the benefits definitely outweigh the costs!benefits definitely outweigh the costs!

Page 28: Snakepit or Shangri-La?

For more information, please contactFor more information, please contact Jessica atJessica at [email protected]@rochester.edu

or Kathleen ator Kathleen at [email protected]@rochester.edu

PowerPoint presentation available atPowerPoint presentation available atwww.rochester.edu/provost/ir/www.rochester.edu/provost/ir/

Thank you!Thank you!