Implementing a DW in parallel with a new application: How not to go nuts!

Presentation Summary

Kimberly Griffin,The University of Chicago, Dan Barrett, Sr. DW Developer

Building a data warehouse for a source system already in production is challenging enough; the job becomes even trickier when the source hasn’t even been configured, let alone built! We'd like to share some lessons learned from both the BI and DW perspectives during the planning, design and build of a data warehouse in conjunction with the implementation of a new application system (in this case Click Commerce Grants and IRB modules). The presentation will include:-Planning:• Fight for resources!• Coordinating the project plan and milestones (and explaining why reporting needs to lag behind)-Design:• Strategies for piggybacking application and reporting requirements gathering• The art & science of reading functional specs from a data warehouse point of view• Incorporating reporting needs into the application design• Coping with an ever-changing transaction data model and inability to conduct data profiling, and the impacts on source-to-target mapping-Build:• Building a BI layer on top of an empty database• The extract – easier to manage when the DW team writes it• Maximizing transform & load coding before extracts are available-Test:• Planning ahead with the application team to coordinate test data that meets reporting needs• Ensuring that test data is delivered in sufficient time for reporting to test thoroughly-Manage: • Tips for integrating a reporting team into a bigger project team