Herding Effective-dated Cats Challenges for a Student-term Data

Herding Effective-dated Cats Challenges for a Student-term Data

Herding Effective-dated Cats Challenges for a Student-term Data Model in a Decentralized Business Environment Jonathan Havey Information Analyst SUNY Buffalo Office of Institutional Analysis

Successful 2011 Implementation of PeopleSoft Campus Solutions On time and Under Budget

Campus Community Admissions Student Records Academic Advising Student Financials Financial Aid

Reporting was Not Ignored Blackboard Analytics (iStrategy) was purchased to meet two immediate needs: o Instant data warehouse o BI Dashboard A pre-existing OBIEE instance, utilized for financial and human resources reporting, was leveraged to deliver

web-accessible, canned reports based on the BbA generated dimensional models A custom reporting app for State Reporting was developed in-house Full Speed Ahead Data Validation was promising o Counts were within range

o Spot-checking revealed no obvious errors o BbA-generated dimensional models revealed no major differences from PeopleSoft source tables. For example, counts of students by Career and Term in the Student Term model mirrored counts of students by Career and Term in PS_STDNT_CAR_TERM OBIEE reports fit the bill

Change Management: The Reporting Community In anticipation of reactions to the new data warehouse, a group of reporting professionals from our many decanal areas was convened These DDAs, or Distributed Data Access professionals were to be trained in the new data models Our legacy warehouse was popular and easy to use, as

any query of the strongly de-normalized views would never find more than one student record per term PeopleSoft, much richer in its dimensionality, a fact reflected in the BbA-generated models, was not so simple On the Movement of Cheese The DDAs, together with the users of the newly minted

OBIEE operational reports, found themselves needing to deal with the possibility that a student might have more than one record per term in the new models These changes were managed through educational efforts As time went on, however, some legitimate holes in the cheese were identified

Associating Effective Dated Rows to Student Terms PeopleSoft handles enrollment data with a StudentCareer-Term concept. Career/Program/Plan (majors, to the rest of us) data is effective dated. You are not a major as of a certain term; you are a major as of a certain date. BlackBoard does a good job of associating effective-dated data to terms, but it needs something to work with: A student term record, first of all, and an effective dated row

that falls within those term dates. Our validation was done against converted data. After golive, ~200 functional users able to edit data started throwing curveballs at BlackBoard. Challenges arose Challenge 1: Disappearing Applicants The University at Buffalo receives approximately 500 Early Decision Applicants every year. Nightly bulk loads of SUNY and Common Application data

are processed, assigned IDs, and imported into PeopleSoft. Students indicating an interest in Early Decision are automatically added to a Student Group, which is an effective dated action. Inactivations are also dated. We get a high yield from Early Decision applications, but not as high as the sample OBIEE report below indicates. Between

November and April, we lose a third of our Early Decision applicants, which makes the yield look higher than it is. Where did they go? Oftentimes, Early Decision Applicants will change their minds and request to be considered for regular admissions. When they do so, their membership in the Early Decision Student Group is inactivated by adding a new row:

An additional source of inactivations in the Early Decision Student Group is incomplete applications at the time of the Early Decision deadline. These applicants are rolled into the Regular Decision pool. Data mining revealed that the gap between our Early Decision totals captured in the Fall and the numbers we were

seeing on our report in the Spring were due to the dating of the Early Decision Student Group inactivations. Possible solutions included: Changing our effective dating practices (major business process change for admissions office) Customizing the Student Groups model to capture any Student Group membership falling out of rangemany undesirable effects

Pointing the OBIEE report to source data for Student Groups Challenge 1: Solution This particular challenge was met by creating a new Historical Student Groups dimension in the OBIEE .rpd. For a given Student Group membership, an Activate and Inactivate Date is calculated. Using these columns in expressions allows a true date-to-date

comparison of Early Decision applicant totals, no matter when the inactivation happened. (Each row of dates below belongs to an individual applicant.) Here, we bypassed the logic in the BbA models by using the PeopleSoft source table. Downside: This solves the problem in OBIEE only. Challenge 2: Applicants that wont Go Away The rise in electronic application submissions has created

opportunities and challenges for colleges and universities. The University at Buffalo receives tens of thousands of applications every year from the Common Application and from a similar SUNY electronic application form. There is a tendency for applicants to use the strategy when in doubt, check every box. In our case, many more students selfidentify as EOP (Educational Opportunity Program) than actually qualify. As with Early Decision, we use PeopleSoft Student Group functionality to track this cohort, and

inactivate non-qualifying candidates. Solution to Challenge 2: Business Process Change The solution to challenge #2 was to alter business process and create a new EOP Student Group that is distinct from the current group, which is populated at admissions. The student EOP Group will contain validated members in good standing rather than simply

self-identified candidates. Achieving this solution required bringing two offices together to devise and agree on a solution. These can be difficult situations to negotiate. Challenge 3: Disappearing Graduates The University at Buffalo enrolls approximately 30,000 students in Undergraduate, Graduate, Medical, Dental,

Pharmacy, and Law Careers. The conferral results here, for June 1, 2014, represent less than 10% of the expected total graduates. Examining the records of the few students who were captured here revealed they were all early conferrals. Which Way did they Go? The effective dating of the Program Completions after the end of the Spring Term might lead you to think that these

graduates are being incorrectly associated with the Summer Term immediately following. Once again, however, the totals do not approach the expected results for either term separately or in aggregate: My Kingdom for a Term Exhibit A: Sam Students Program/Plan Stack

Exhibit B: Sam Students Term History The effective-dating practices (historical in origin) are at odds with term activation (which is set-up to help control postconferral registration). The negative effects are strictly in the reporting arena, however. Possible Solutions to Challenge 3 Just use Completion Term or Expected Graduation Term

o Might work here, but term dimensionality like this is generally unavailable for program actions Effective date differently o Major training/oversight/coordination challenges Customization to existing views o Technical resources scarce Build a new model o Begs the questionwhy have the view in the first

place? Two Solutions were devised: A Customization and a New View Customization Specs Find students final terms in a Career Find the end-of-term dates for those Terms/Careers Look for Program Actions effective dated after those terms End-of-Term Dates

Define new fields that display the maximum effective-dated Program Status after a students last term Two Solutions were devised: A Customization and a New View Customization: 4000 Missing Graduates Found Of Note: The truncation of individual students Program Plan effective-dated history is especially noticeable in the cases

where a student who had previously been on leave or discontinued due to a gap in enrollment returned to graduate. Two Solutions were devised: A Customization and a New View New View: Fuller Detail for Reporting on Degree Applicants in addition to Awardees Delivered suite of BbA data models does not include detail on degree applicantsno pipeline

Degree Awards models truncates number of awards Here, we found a custom view was called for Challenge 4: The Applicant Made out of Rubber Some savvy functional users pointed out that some of our new students had GPAsnot an expected result except in the case of some non-degree-seeking students with previous work.

This students Admit Term of Fall 2014 together with certain upstream admissions attributes caused them to appear briefly as a New Freshman in Fall 2014, when they had in fact started in Fall 2012. Solution to Challenge 4: Fix the Data In this case, a customization, .rpd update, or business process change was insufficient. We needed the functional users to stop making these changes.

Edit checks are customizations in PeopleSoft. We decided to utilize some custom functionality that our technical team had recently provideda Data Integrity Checker that would utilize built-in queries designed to identify particular errors after they were made. Data Integrity Checker When report-based solutions, business process redesign, and data model

customizations are too expensive, you have to go to the source: The data itself. Here, the work is in the query design. It has to be a surgical instrument to ensure the trust of the users that they are part of a successful process. When the query runs at night, it willif records are found email the alias with a

link to the query output, every morning until the records are fixed. Sample output from The Data Integrity Checker. Attempt is made to cue the functional user in to the nature of the problem. Here, the Data Integrity Checker is finding Minors that are coded incorrectly.

Summary Systems that rely heavily on effective dating can create challenges when attempting to associate that data to a higher level of aggregation such as a term. Our instant warehouse, a good product, was able to handle most of the effective dated data in our implementation. The University at Buffalo has nearly two hundred users

across campus performing effective dated transactions. Training and oversight is an exercise in herding cats. After using Business Process Redesign and customizations to dimensional models, we are now focusing on data quality, making sure it is within expected ranges.

Recently Viewed Presentations

  • İhlas Ve Ni̇yet

    İhlas Ve Ni̇yet

    Bu mükâfât 700 misline kadar çıkar. Eğer yapılan iyilik Allah Teâlâ'nın çok değer verdiği davranışlardan biriyse, kul da o işi ihlâs ve samimiyetle yapmışsa, mükâfâtı 700 misliyle de kalmaz; hesabını sadece Cenâb-ı Hakk'ın bileceği daha yüksek ölçeklerle ...
  • Larry Rubio Chief Executive Officer Riverside Transit Agency

    Larry Rubio Chief Executive Officer Riverside Transit Agency

    City of Beaumont. City of Calimesa. City of Canyon Lake. City of Corona. City of Eastvale. City of Hemet. City of Jurupa Valley. City of Lake Elsinore. City of Menifee. City of Moreno Valley. City of Murrieta. City of Norco....
  • 1èrePARTIE: ORGANISATION DES EXAMENS

    1èrePARTIE: ORGANISATION DES EXAMENS

    Implication personnelle dans le développement du judo jujitsu justifiée par le candidat par attestation délivrée par le CORG et attestant d'au moins un titre ou une fonction depuis son dernier grade parmi : - Enseignant en exercice, - Commissaire sportif...
  • Principles of European Tort Law A Critical Examination

    Principles of European Tort Law A Critical Examination

    Principles of European Tort Law A Critical Examination Ken Oliphant (Cardiff Law School) BIICL, 23 November 2005 1. Why Principles of European Tort Law? Full harmonisation (a European Civil Code) Other "hard" harmonisation "Soft" harmonisation Development of a distinct EC...
  • The Italian Campaign - Weebly

    The Italian Campaign - Weebly

    Background German and Russian Troops German troops faced disaster in Russia. They were unable to handle the freezing Russian weather. The German army surrendered in 1943. The Russian forces now turned their attention towards Germany. After failing in Russia, Hitler...
  • O papel do controle social e a nova

    O papel do controle social e a nova

    A lei 13.019: destaques. OSC (site) nome e CNPJ. data, identificação da parceria e órgão. descrição do objeto. valor total e valores liberados. situação da prestação de contas
  • Relationship Of Organizational Identity and Organizational ...

    Relationship Of Organizational Identity and Organizational ...

    Relationship between organizational identity and organizational performance. Proposition 1: Since organizational identity impacts organizational goals and strategies, the degree to which an organization identifies with its normative versus its utilization identity will impact the goals and organizational performance measures the...
  • EGEE Middleware

    EGEE Middleware

    Mixed Attributes ¤ Association between one CE and one or more SEs (objectclass: GlueCESEBindGroup) • GlueCESEBindGroupCEUniqueID: unique ID for the CE • GlueCESEBindGroupSEUniqueID: unique ID for the SE Utilisation du Système d'Information Services Grille Matchmaking Le SI est interrogé par...