Tuesday, May 1, 2012
01:30 PM - 02:20 PM
A variety of approaches have been used to assure availability of data for Business Intelligence. Intel's approach to fully integrate their data is to have a central Data Warehouse linking to data structures that allow data easy traversal. However, specific use cases are handled using dimensional models. The co-existence of these two modeling techniques, entity-relationship and dimensional, is not always harmonious. In successfully following this path, we found a number of surprises, from what to do with character sets to how best to transmit the data. This presentation shares some of these surprises, explains how we worked with them, and gives reasons for our solutions.
- The Business Intelligence capability you are trying to build, moving from reports to advanced analytics
- Why a solid data foundation requires data management governance to be in place
- Entity-Relationship Modeling vs. Dimensional Modeling
- Introducing Reporting Anomalies
- Flexibility vs. Tuning for particular use cases
- Extra data analysis required to understand divergent data sets
- Some unexpected hurdles we encountered and how we addressed them
- Building for the future of Business Intelligence Analytics
Peggy Schlesinger is a Data Analyst at Intel Corporation and a Certified Master Enterprise Architect. She has over 25 years experience in various aspects of data management; has worked in a number of corporations using a variety of methodologies, has a Masters in Business with an MIS concentration, and has published several papers presented at technical conferences. Peggy has an appreciation of the practical aspects of implementing methodologies that are complete and support the development activity.