Information Architectures

moderated by Laura Cisneros and Mike Torregrossa, University of Arizona

The session opened with the participants introducing themselves and describing the Information Architecture issues of particular interest to them and their institution. The issues and questions varied widely. It was apparent that at most institutions the pursuit of an information architecture is still in its infancy and that it will be some time before actual experience will prove the worth of the pursuit. Until such time we want to know if we share an understanding of the components of an information architecture, where it fits in the organizational structure, and, how it facilitates integration.

Some of the questions and concerns posed during this session and the points made by the participants included:

• Who will use the information architecture?
  Anyone developing a data warehouse, or, developing or purchasing a new administrative system would want and need the information provided by the Information Architecture. One participant commented that they are now requiring vendors to provide a logical data model of their systems. By comparing the institutions Information Architecture to that of the vendor's they can quickly identify how well the vendor proposal will meet the data requirements of the institution. At the time of the conference they had yet to receive a logical ERD from a vendor so whether their request will be honored is not known but it was agreed that if more institutions started asking for logical models that the vendors would make them available.
• What does information architecture resolve?
  An Information Architecture is a logical view of the business. If properly executed and implemented it will resolve unplanned data redundancy, data integrity issues, and, duplication of effort.
• How are institutions delivering metadata?
  Metadata is being delivered in relational tables within the data warehouse environments. One institution has an institutional data catalog application where institutional metadata is stored. As data warehouses/data marts are developed the metadata is loaded from the catalog into the appropriate database. The data catalog is the central store for institutional metadata and is available for anyone needing descriptions and definitions.
• Users must be educated on report content and interpreting information.
  As more information is made accessible the need to educate users increases. The University of Pittsburgh captures the questions posed by users and basis the content of their training on those questions. Everybody agrees on the need for training, however training large numbers of users with diverse needs is a major undertaking. Data quality improves through its use and understanding.
• How does an information architecture support integration?
  Integration requires a clear understanding of the business and the data required for it’s operation. One of the components of an information architecture is a map of the activities that support the business, the subject areas involved in the activity, and the attributes of each subject area. Relationships are noted between subject areas. This mapping, often referred to as a logical entity relationship diagram (ERD) is a stable statement of the business and how it operates. This information which changes little over time becomes the basis for integration.
• What is the difference between information and data?
  Data are the components of information. Information is data assembled to form a meaningful message for the user.

Other questions and concerns raised but not discussed in much detail:

• Managing information and data must be separated from technology.
• Do most institution have a central Data Administration?
• Who has an Information Architecture that can be shared?
• How many employees are assigned to Data Administration support?
• Policies are needed to support Data Administration.
• How are data access and security being addressed in a data warehouse environment?
• How is data quality enhanced by an Information Architecture?

Summary

Data Administration and Information Architecture are still in their infancy at most institutions. At this point in time there are more questions than answers. There was some agreement to continue this dialog on the electronic discussion group DASIG.