Metadata-driven load processing at Seattle University

Presentation Summary

Eric Larsen -Seattle University

This presentation gives a technical overview of the design and processes around the metadata-driven load processor used by Seattle University's dimensional data warehouse. Metadata-driven processes accelerate integrating new data as well as implementing new dimensions and fact tables. This also decreases the amount of development required to implement new data warehouse objects. As part of load processing, in-depth data validation is also metadata-driven and allows for targeted reporting of data errors to the appropriate offices on campus.

Currently, the load processor manages both the staging and reporting schemas that encompass 280 staging tables, 110 dimensions, 130 fact tables, and 87 custom script tasks. 4 major systems are integrated with the ERP being the primary source. The data warehouse has been fully live since May of 2008.