 |
 |
 |
What is the University of Maryland Data Warehouse?
|
|
|
|
|
The University of Maryland Data Warehouse is a collection of integrated institutional data which can be accessed by authorized UM business managers, administrators, service providers and institutional researchers for the purpose of performing analysis, producing ad hoc queries and reports, and maintaining data subsets. Data are extracted from the campus business transactional systems, put into the data warehouse, and retrieved and analyzed at the user's desktop. If you would like a demo of the data warehouse or more information, please contact Barbara Hope on 301-405-6819 or Lori Kasamatsu on 301-405-1713.
The following information is covered below:
Data Warehouse Sponsorship
The University of Maryland Data Warehouse is a joint project among Administrative and Enterprise Applications (AEA, including the Office of Data Administration, Administrative Enterprise Application Support Services and Student Applications Services), the Office of Institutional Research and Planning, campus data stewards and managers, and end users of institutional data. All of these groups work as a partnership to provide the hardware, software, training, support and data subsets.
Relationship of Warehouse to Transaction Systems
Data used by the University of Maryland Data Warehouse are extracted from transactional systems which handle the day-to-day operations of the university -- collecting, processing, and storing operational data. Examples of transactional systems are the Student Information System (SIS), the Academic Resource System (ARS), and the Financial Resources System (FRS). Transaction systems handle processes such as registering a student, approving an appointment and posting a journal voucher. These transaction systems were not designed to provide managers with the information they need to conduct trend analysis, particularly when data are located on different platforms. Data warehouses, unlike transaction systems, are specifically designed to handle query capability and analyses.
How the Warehouse is Built
To build the University of Maryland Data Warehouse, data were identified for inclusion (broad range focus groups met to identify data needed), "cleaned" (checked for redundancy, data errors, conflicting definitions) and restructured (data naming standards applied and elements grouped for ease of querying). In addition, "derived" data were added. Derived data are newly created data elements which are computed from a combination of several existing data elements or by applying computational formulas. An example of a derived field is "age" which is derived from a person's birth date. The process of identifying, cleaning, restructuring and deriving data is followed for every data subset (student, personnel, faculty, etc.) identified for inclusion in the data warehouse.
Once the data have been identified through the above process, programs extract the data from the transaction systems and put them into the data warehouse. Users can then access data directly from the data warehouse and run canned and ad hoc queries as desired. Because they are extracted from transaction systems, data in the data warehouse are "view only" and cannot be updated. Any changes and/or corrections to the transactional data must be made in the actual transaction system and are reflected in the next refresh cycle.
Data Refresh Cycles
After data are initially loaded into the data warehouse, they must be refreshed on a periodic basis. Refreshing of a data subset is determined by a number of factors including, static/non-static nature of data, need, etc. Some data subsets are refreshed several times a week and others may be refreshed once a term.
The following schedule reflects current refresh cycles in the University of Maryland Data Warehouse:
- Faculty Appointment Data - Nightly
- Faculty Other Data - Nightly
- Financial Aid Data - Wednesday and Friday nights
- Name & Address Data - Friday nights
- Personnel Data - Nightly
- RAA Summary & Detail Data - Monday, Wednesday and Friday nights
- Scheduling Data - Monday, Wednesday and Friday nights
- Student Data - Tuesday and Friday nights
- SIS Utility Data - Nightly
It is important to be aware of the refresh cycles when running a query. The "delay" between the actual transaction system and the data warehouse refresh cycle can affect the results obtained in a query. Refresh dates can be checked for each data subset within the data warehouse via header tables or by checking the extract date associated with each record within a table.
Data Warehouse Availability
The University of Maryland Data Warehouse is available for use by authorized users 24 hours per day from 7 AM Monday until 1 AM on Sunday.
|
|
|
