University of
ConnecticutEnterprise Data Warehouse
University BI Services
Contents:
About the Data Warehouse
The Data Warehouse is a repository of University data from campus transaction systems. Some examples of transactional systems are Student Administration System and the Financial Records System. Data is extracted from the campus transactional systems, put into the Data Warehouse, and it is then accessed and analyzed from the customer's desktop.
The Data Warehouse is built from the transactional systems. Data is identified for inclusion, checked for redundancy, and restructured if necessary (using a combination of fields to create a new field). Once the data is identified, it is put into the Data Warehouse. The data is periodically refreshed, determined by the static/non-static nature of the data, need, etc. Some Data Marts are refreshed nightly, others not as frequently. This data is for queries only; any changes to data must be made in the transactional system.
Existing Data Marts – BDM, SADM, Focus files
Existing Data Marts will continue to exist until their need is replaced as the data is moved into the Data Warehouse.
Our Data Analytics Vision
Improve the organization’s capabilities by providing business insight to those that need to know to make quick and accurate decisions and facilitate external reporting obligations.
- The right information
- The right time
- The right format
Background & Overview
In the past decade, institutions of higher learning have recognized an increasing need for data and information as the basis upon which to direct the large and complex universities of today. Traditional transaction logs, operational listings, and period-ending result reports are still crucial to successfully conduct the ongoing operation of the university. However, new elements of data, new “views” of data that transcend the university’s usual operational structures, and new tools for accessing and reporting that data are increasingly required components within an organization’s data architecture and information services. Such data is specifically designed to fulfill the need for more analytical types of data studies, “what if” modeling projections, trend studies and projections, correlations of data within more specific subsets and demographic scopes, and identified performance metrics.
Addressing the five goals of the academic plan: A number of the metrics for graduate education and research would be very valuable for department heads and school/college Deans; particularly enrollment figures, number of degrees awarded, research awards, and research expenditures. Successful research and graduate education programs are huge drivers in moving a University higher in national rankings, and being among the top-20 public universities in the nation is a stated aspiration in the Academic Plan. Being able to measure the success of particular programs depends on accurate metrics.
Definitions
The terms Data-Mart and Data-Warehouse have been used at UConn to represent the same thing. We offer the definitions below for clarification during our discussion.
Transactional Data Store
- Online Transactional Processing (OLTP)
- Live transactional data from online systems
Operational Data Store (ODS)
- Aged transactional data from online systems
- Often a copy of the Transactional Data Store
Data Warehouse
- Structured data from online systems used for On Line Analytic Processing (OLAP) stored in a repository that houses data from multiple data sources dealing with multiple subject areas.
Data Mart
- Structured subset of data from online systems used for OLAP and is oriented to a specific subject area. This implementation may be a logical subset of the Data Warehouse or a physically separate implementation.
Structured Data
- Structured data is generally comprised of Dimensions and Facts arranged in tables and organized in a relationship
- Dimension: Conceptually, a business process or object, a thing to inquire about.
E. G. Enrollment - Fact: Raw facts and aggregations (in this case, about enrollments).
E. G. Average semester enrollments for a regional campus
- Dimension: Conceptually, a business process or object, a thing to inquire about.
Architecture
