Download a PDF of the Session Descriptions (updated 2/29/2008) HEDWpresentations.pdf

9:15 to 10:15 am - Break-out Sessions

Latham D-E-F
Change Data Capture Made (sort of) Easy, Mike Koontz & David Melton, University of Richmond

Data warehouse environments generally rely on source system processes to determine the deltas ( inserts, updates and deletes) within the database. Without source system generated timestamps, determining these changes rely on traditional methods such as batch compares or custom database triggers. Oracle's Change Data Capture (CDC) feature offers the ability to read the database transaction logs to determine the deltas without the heavy load processing or invasive customization normally associated with traditional methods. University of Richmond has implemented CDC with a set of custom written PL/SQL procedures and metadata tables to automate the capture, transformation and transfer of the deltas into our data warehouse. This presentation will focus on the particular CDC method Asynchronous Hotlog, the rational behind choosing this method and the various additional steps required to make the data useful for data warehousing.

Solitude
Data Warehouse and BI Approved: Now what?, Joseph Kerr, University at Buffalo

The presentation will describe an overall framework for designing, building, and deploying a data warehouse and business intelligence solution using the UB experience as an example. It will cover stakeholder and vendor management, project planning, roles and responsibilities, designing and building data models, reports and dashboards, and deployment. It will identify the most significant challenges, lessons learned, and the hidden value these projects deliver. There will be a short demo of the UB solution as well.

Assembly Hall
Data Governance at MIT, Mary Weisse, MIT

MIT has a very comprehensive infrastructure that has helped make our life a little easier when it comes to Data Governance. We are really good at managing data, authorizations, and authentications. There are still areas where we can improve (local data sources, real-time availability of data, etc). We'd like to share what we've done, where we are, and have an open discussion about where we and others are headed.

10:30 to 11:30 am - Break-out Sessions

Latham D-E-F
From Data Warehouse to Business Intelligence: The Michigan Journey, John Gohsman & Sean Mallin, University of Michigan

The University of Michigan has been building its relational data warehouse since the early 1990's. Data was exposed to end users via several data sets that could be integrated. An ERP implementation required the University to rebuild new versions of these data sets. After the ERP system was stabilized, executive officers found that we were not leveraging the vast amount of data to make better decisions. In response, Michigan under took the development of a Business Intelligence (BI) strategic plan. During the plan development, we realized the need to push our data warehouse and BI to a higher level which required a more sophisticated technical architecture and advanced data models. These advancements would improve query performance, ease access to complex information, and improve analytical capability. We believed the results would be increased user adoption and improve decision making. As part of implementing our BI strategy, we selected a delivered BI solution to augment our in-house developed solution. Campus energy is now up and demand for information is high. We will share the Michigan journey, demonstrate how a delivered solution jump-started our BI plans and demonstrate the HR product we co-developed with iStrategy.

Solitude
Open Source Reporting at Indiana University, Robert Serbent & Rebecca Gribble, Indiana University

This presentation will talk about the history of delivering decision support data at Indiana University, our current strengths, needs and deficiencies, and our future vision for the use of open source reporting coding projects and business intelligence software at IU.The Open Source project we chose to start with was the Eclipse based BIRT project. Eclipse is an open source community whose projects are focused on building an open development platforms with extensible frameworks, tools and runtimes for building, deploying and managing software systems through their lifecycle. BIRT stands for Business Intelligence Reporting Tool. It is an official Eclipse project that provides a reporting system for web applications, especially those based on Java and J2EE. BIRT consists of two major components: a report designer based on Eclipse, and a runtime engine that can be added to just about any application server.Our talk will discuss why Indiana University choose to implement BIRT, demonstrate its best features, both from an end user perspective and as a report developer, describe its architecture, and discuss how we were able to integrate it with the only outlay being a minimal number of labor hours. We will also focus on how we see open source products being used in BI at Indiana University in the future.

Assembly Hall
Building A Solid Foundation for Data Quality and Data Integration, B.K. Chen, University of Washington

As more and more data from different data sources (HR, Academics, Financial, etc) are loaded into university BI environments, data quality and data integration issues become extremely challenging. Master Data Management is a strategy taken by UW DSS to tackle these difficulties. One key component towards facilitating data integration and the creation of Master Data in UW DSS is the development of a standard time dimension.

Higher education institutions run on many different calendars. In UW's case, there are academic, federal, state fiscal calendars, etc. Operational systems typically have defined and built their own ways of recording event times. The users of data, however, often require reporting across different data sources and business domains. This presentation will discuss how this standard time dimension, as one of the basic steps of building master data, was designed, developed, and implemented. I will use that case to explore the implication of Master Data Management in higher education setting.

11:30 am to 12:45 pm - Latham A-B - Lunch

1 to 2 pm - Break-out Sessions

Latham D-E-F
Empowering Institutional Change Across Virginia's Community Colleges Through Data: Moving Toward a Comprehensive Data Warehouse Through Two Datamarts, Bruce Bartek, Tim Jones, Susan Wood, Monty Sullivan, Virginia Community College System

The system office of Virginia's 23 community colleges is leading a comprehensive data warehouse project. Providing impetus for this project was the conversion four years ago of the legacy student information system to a PeopleSoft/Oracle SIS at which time most of the standard reporting functionality built up around the legacy system was lost. Additionally, the introduction of a new SIS with increased functionality into a multi-institution environment created many data quality issues. This session will describe the progress to-date as the system has moved from limited access and inconsistent data towards an integrated portal. A curriculum datamart focusing on enrollment, awards, and program productivity data was developed and rolled out to college users at the "researcher" and "consumer" levels. Nearing completion and rollout is a second datamart focusing on retention that enables the user to define and track a cohort on a number of analysis measures. These two datamarts represent proofs-of-concept whose use by the colleges will inform a strategic value assessment. The full build-out of the data warehouse will incorporate outcomes from the strategic value assessment. In this session, presenters will describe the evolution of the project, involvement of key stakeholders, challenges in development and implementation, and lessons learned.

Solitude
Tackling the Challenge of Integrating Locally Developed Applications with University Data, Lance Tucker & Josh Millinger, Boston College

The proliferation of desktop database technologies over the last twenty years has resulted in a myriad of data quality and security issues on campus. This presentation will focus on Boston College' strategy of using Oracle's Application Express in conjunction with our Oracle Data Warehouse to deliver safe and efficient information delivery solutions to our customers. Administrative groups have been able to build their own local applications without IT assistance. Applications like FileMaker Pro spread quickly. Users became proficient with this technology in a short period of time and built solutions without any IT costs. Many useful applications evolved that serviced the University. However, key problems also developed around data quality, data security, and data management. This presentation will focus on Boston College's effort to address the two largest issues these local applications presented:

• Multiple Systems of Record
• Data Security - authentication and authorization

The presentation will review the architecture we chose to address these issues. We will go into detail on the data delivery and security mechanisms used to provide a more robust solution. We will also explore the policies required, business processes, and associated costs to support this solution. A working application will be demonstrated using this approach.

Assembly Hall
Lessons in Faculty Activity: From Politics to Analytics, Aaron Walz & Beth Ladd, University of Illinois

One common challenge for academic administrators is understanding the range of activities faculty are involved in. This information is essential context for evaluating research productivity, making promotion and tenure decisions, determining where to best allocate scarce resources, and responding to surveys. However, it often remains elusive because it requires integrating data spanning finance, pre-awards, grants, payroll, human resources, and student enrollment. The business questions that must be answered to create individual puzzle pieces, as well as the complex logic required to connect them, combined with culture, organizational and political factors create real challenges when trying to establish common business rules and definitions around such questions as "who are my faculty?" This presentation will describe a project at the University of Illinois that designed and built an OLAP cube to support Faculty Activity Analysis. It will begin by covering the many political and organizational challenges encountered in starting up and sustaining such a project and how they were addressed. It will also describe the overall approach, design challenges, business questions and how they were resolved. Finally, it will demonstrate the end result and describe the reaction from the intended customers.

2:15 to 3:15 pm - Break-out Sessions

Latham D-E-F
Building Next Generation Data Marts at Cornell, Jeff Christen, James Singleton & Yiorgos Marathias, Cornell University

Cornell University has chosen and is implementing a path towards an Enterprise Data Warehousing solution. This strategy involves:

• Using the Kimball Methodology to manage the project lifecycle along with developing Dimensional Models (STAR Schemas) for new Data Marts;
• Utilizing the mature infrastructure and resource with Cornell Information Technology;
• Utilizing both Internal Resources and an External Consultant Organization for new data marts and when re-engineering existing data stores;
• Delivering data marts in support of new Operational Application roll-outs. Specific areas that will be expanded upon within the presentation include:
• Business User Requirements definition and translation;
• Project Planning, Scope and Commitment;
• Data Warehousing Development Methodology;
• Cornell Universities Data Warehousing Technical Architecture;
• Utilizing Source Applications / Legacy and Local Data.

Solitude
The Virtues of Cross-Training, Steve Grantham, Boise State University

At Boise State University we are in the process of implementing the iStrategy Data Warehouse for Student data. After having worked in the Enrollment Services area on our PeopleSoft implementation for about 8 years, I moved over to our IR department to lead the Data Warehouse implementation. One thing I have found in making this transition is that many of the precise details - some might say idiosyncrasies - of the business practices of offices like Admissions and the Registrar have not necessarily been well known or understood on the IR side of the house. And my transition has driven home to me the point that such offices naturally design and implement their business practices in a way that optimizes transactional processing, not necessarily reporting. I believe that my familiarity with many of the details of the transactional system has made it easier for me to analyze and understand the logic needed to populate the data warehouse correctly. This presentation will give several specific examples intended to illustrate this general principle and highlight the importance of communication between the transactional and reporting units of a university.

Assembly Hall
Cooperative Data Quality Checking and Improvement, Dylan Cooper, University of Arizona

Bad data, or the suspicion of it, is a major concern for reporting and decision support systems. At the University of Arizona, there has long been anecdotal evidence of data quality issues. This has led to problems for data analysts and uncertainty about the reliability of the data in general. During the past two years, the Information Warehouse Office has partnered with the users of the personnel and student information systems to find and correct data quality errors. We designed and implemented a system that checks the data as soon as it is loaded into the operational data store and displays any errors found in a web site. The administrative applications staff uses the web site to correct the data in their systems. This process has allowed us to find and correct thousands of cross-system inconsistencies, missing values, miscoded attributes, invalid dates, etc. Additionally it has allowed us to measure the quality of our data in ways we were not able to before. Happily we discovered that it is better than some people feared. This talk will outline the technological and sociological approach we have taken, the problems we have faced, the successes we have had, and our future plans.

3:15 to 3:45 pm - Afternoon Break

4:00 to 5:00 pm - Break-out Sessions

Latham D-E-F
Tips and Tricks with SQL Queries, Michael Wonderlich, University of Illinois

This presentation will focus on intermediate and advanced techniques in SQL writing. Techniques for analytical processing and advanced relationship combinations will be reviewed. Also tips for writing more efficient SQL will be provided.

Solitude
Enterprise Reporting at the University of Delaware, Kat Collison & Karen DeMonte, University of Delaware

In the fall of 2004, the University of Delaware began converting its Student Administration system to PeopleSoft. . Institutional Research was asked to lead the reporting effort during the implementation phase using a new reporting system, Cognos. PeopleSoft, like other transactional data base systems, is designed for processing information, not reporting. With a system no one knew and a database structure hostile to reporting, progress was slow. A data warehouse was the solution for working with conflicting data structures (the PS system is a transactional database and Cognos prefers dimensional data structures). Institutional Research is now in partnership with Information Technology to design and manage the University of Delaware Enterprise Warehouse - UDEW IT. This presentation will delve into the design and content of the data warehouse from both a functional and technical point of view as well as review its Business Intelligence solution. Emphasis will be on: 1) Gathering specific reporting requirements and business rules; 2) Building the Student Data Mart using Cognos ETL tool - Data Manager, and Framework Manager, 3).Demo of generating reports using Cognos ReportNet and the Student Data Mart.

Assembly Hall
Securing Data in the University of Washington Enterprise Data Warehouse, Anja Canfield- Budde and Glenn Pittenger, University of Washington

The goal of this design is to implement row level and column level security on specific SQL Server objects based on Roles. Based on certain parameters such as organization type, users are classified into a finite set of Roles, which ultimately govern access to sensitive information like Social Security Numbers, Date of Birth, Race or Ethnicity. Within a Role, users receive Span of Control values according to certain parameters, such as Payroll Unit Code (PUC). The information about Roles and Span-of-Control is captured in the campus-wide authentication system ASTRA and will populate to the Data Access Control (DAC) databases from there.

Our talk will introduce the different players in the game: the Roles & Access Matrix, ASTRA, Group Directory Services, Windows Active Directory, and Microsoft Reporting and Analysis Services (SSRS/SSAS). By walking through the design of the SQL server components of the Data Access Control (DAC) schema, our talk will then focus on how we apply security at the data level. We will describe relevant user relationship objects, show entity relationship diagrams, and explain the logical types of tables in the Data Access Control schema.

5:00 to 6:00 pm - Latham A-B - HEDW Business Meeting

7:00 am - Foyer - Continuous Break

8:00 to 9:00 am - Break-out Sessions

Latham D-E-F
Dash and go, working with Data Warehousing…, Jacqueline Nottingham & Pamela McAlexander, Virginia Tech

Because of the variety of customers it must work with, Admissions is considered multifaceted and complex. At Virginia Tech, we have developed business processes for focusing our attention on communicating admissions information internally and to external constituents at the University. This session will provide background on the need for developing a more comprehensive plan for information distribution and establishing stronger relationships with our colleagues, utilizing a data dashboard. We will discuss how Virginia Tech has integrated the use of a Dashboard to create a streamlined process for informing program areas and administrative personnel across campus of needed information. The ability to provide more efficient and higher quality service will be covered as well. This session would be helpful to anyone in the Recruitment or Admission arena. We believe this forum will provide an opportunity for others to learn from our experiences and to help develop their own plans for working smarter with their constituents. It will also provide a structured environment to share and discuss new ideas with others who face similar issues.

Solitude
The University of Texas' project Information Quest Working Model of Data Governance, Jarrett Cole & Darin Mattke, University of Texas

Project IQ is able to deliver exceptional reporting-related products via the implementation of a highly sophisticated and integrated functional model of data governance. UT's team of Data Architect, Data Quality Analyst and BI Developers will diagram and explicate the intricacies of the fine-tuned mechanisms utilized to achieve the delivery of one version of "Truth" University leaders have come to depend on in making the most informed decisions possible. Examples of how the model operates will highlight the roles of Data Stewards, the ETL Process, DBAs, Data Quality Analysts, BI Development, and our Management & Support section. This presentation will also illustrate the procedural processes we define as Logistics, Standards, Deliverables and communication. These areas deal with data staging, data availability, establishing data definitions and ensuring data quality. A successful implementation is communicated via an iterative process of requirement gathering, developing and testing products, and otherwise interfacing with our community of developers and consumers. The tangible outputs are in the form of reports, cubes and metadata repositories. After attending this session, participants will have a road map to follow for a successful data governance model.

Assembly Hall
From the Data Warehouse and EIS to Business Intelligence, the evolution of Penn State, Marta Miguel & Yvonne Riley, Penn State

The presentation will focus on Penn State's evolution from the Data Warehouse and Enterprise Information System (EIS) (Cognos 7) to a University-wide Business Intelligence strategic vision. It will include an overview of the current DW and EIS systems, current governance and stewardship model, and strengths and weaknesses. Both of these systems currently have over 1000 very active users. Then it will introduce the BI Strategic Vision, how we arrived to that vision, how we received leadership approval, and our approach to implementing it. The strategic vision states that information is quickly becoming a strategic asset and that for Penn State to be able to leverage that asset it needs to consider three factors: Governance and Policy, organizational structures, and software, hardware and data infrastructures. Within the strategic vision each of these three factors needs to be re-assessed. The presentation will explain why information is becoming a strategic asset as well as describe the proposed organizational model and reference technical architecture. This presentation will be useful for people trying to justify a Business Intelligence implementation at their universities or for someone trying to create a university wide strategy for BI.

9:15 to 10:15 am - Break-out Sessions

Latham D-E-F
Association of American Universities Data Exchange DW, John Scanlon & Mary Weisse, MIT

MIT has designed, built, and currently hosts a Data Warehouse for The Association of American Universities Data Exchange (AAUDE). AAUDE is a public service organization whose purpose is to improve the quality and usability of information about higher education. Their membership is comprised of AAU institutions that support this purpose and participate in the exchange of data/information to support decision-making at their institution. This exchange is facilitated by the DW that resides at MIT. MIT would be interested in sharing the DW design, the development process, and how we support our customers at these Institutions.

Solitude
Expanding Data Warehouse Services: Opportunities and Risks, A Panel Discussion, Suneetha Vaitheswaran, University of Chicago; Andrea Ballinger, University of Illinois; Ted Bross, Princeton University; Ora Fish, Rensselaer Polytechnic Institute; Jeffrey Glatstein, University of Massachusetts

A panel of colleagues from higher education institutions who have deployed enterprise data warehouse solutions will discuss their alternate perspectives regarding the demand, benefits, challenges, and decision points in expanding data warehouse strategy and reach. Additional services which may be discussed are using the data warehouse as a feed source of data to other enterprise systems, increasing the load or refresh frequency to near real-time, distributing standard report development beyond the central development group, etc.

Assembly Hall
Developing & Using Key Performance Indicators (KPI) to Provide Institutional Strategic Direction, Jack Mahoney & Jeff Stark, Rensselaer Polytechnic Institute

Higher education institutions, like all industries, have numerous measures that provide insight into areas requiring attention. The challenge in any organizational setting is to clearly identify an agreed upon set of business metrics that represents the vital areas of the organization. Rensselaer Polytechnic Institute will share their experience in developing their Key Performance Indicators and Faculty Workload applications, both of which provide essential tools to focus on areas defined by the university as critical and strategic. The discussion will focus on the process Rensselaer used to define a set of business metrics, why these particular metrics were selected, how they were structured in an easy-to-use dashboard, and where they are being used to produce actionable results. Coupled with the presentation will be a demonstration of Rensselaer's Performance Metric and Faculty Workload Dashboards.

10:30 to 11:30 am - Break-out Sessions

Latham D-E-F
Using nVision and PeopleSoft Security with Kimball-style Data Warehouse, Craig Leslie, DePaul University

nVision is a widely used MSExcel based BI tool among PeopleSoft users, especially among financially oriented applications. Furthermore, nVision can combine the results of multiple queries and complex data manipulations on a single page report, taking advantage of the full range of Microsoft presentation formatting. nVision has suffered from slowness in the PSQuery layer when extracting data from complex PSoft transaction tables. But when a highly tuned, dimensional DW is used as the source, performance increased dramatically while retaining the reporting advantages of nVision.

Solitude
Faculty Workload Analysis, A Panel Discussion, Emily Thomas, Stony Brook University; Robert Duniway, Seattle University; Richard Howard, University of Minnesota; Jack Mahoney, Rensselaer Polytechnic Institute; Aaron Walz, University of Illinois

Measuring faculty workload is important to university management and complicated. The panelists in this session will lead a discussion of the analytic, technical, and political issues involved. Questions include: How can data on teaching and non-teaching workload be captured? What metrics should be reported? What governance issues arise?

Assembly Hall
Establishing a Business Intelligence Competency Center (BICC), Michael Wonderlich, University of Illinois

In today's industry the hot topic is a Business Intelligence Competency Center (BICC). Business Intelligence is becoming a critical piece in the success of a data warehouse environment. A BICC provides an organized method to address the staffing, process, BI culture and infrastructure needs of your university. In this session we will

• Define the components of a BICC.
• Examine how existing services match to the components of a BICC.
• Review how to assess the gaps in our organizations
• Discuss ideas how to fill those gaps.

11:30 am to 12:45 pm - Latham A-B - Lunch

1:00 to 2:00 pm - Break-out Sessions

Latham D-E-F
Birds of a Feather---Sessions to be determined by attendees

Solitude
How not to manage a data warehouse development project, Bob Duniway, Seattle University

Seattle U. demonstrated the value of reporting against a central data warehouse in 2003, and we wanted to expand our efforts. Although we are going operational with our new and improved dimensional data warehouse in March, I would be lying if I said the process of developing this solution was well managed by the university. This is a Jesuit university, so lying is not an option. Instead, I'll share the truth of our process, what went right, what went wrong, and how we recovered.