Data Warehousing Interchange
July 12, 2005

Focus: This Interchange focused on the recent technologies and new best practices in developing and managing data warehousing systems.

What is Data Warehousing?
-A process by which to store data
-A repository of data to analyze
-A point of collection of data for the provider
-A place where data can be displayed and shown as a big picture
-Integration of data from multiple systems or sources that standardizes and cleans data to a single source of truth.

Key points to look at when starting data warehousing.
-Buy in from upper management and getting them to define metrics of what to report.
-Quality of data and consistency from IT.
-You must present assurances up front to gain the trust of programmers.
-Standardization of data is key to making it work.
-Trying to put a ROI on data warehousing is very difficult.
-Cleansing data is very important.
-Know user expectation.
-Data Warehouse is just a piece of the B.I. plan, and it is always changing just like the business.
-Try to help individuals out in your organization using data warehousing and let them show upper management what it is worth.
-Don’t try to do too much at one time, start with different pieces and put them all together to build one unit.

Quality Assurance:
-1 QA
-2 business analysis= verification of data (may be automated)
-3 user acceptance= responsible for giving the o.k.
-4 then live data

If the user has sent you bad data, the user should fix data and you should never manipulate the data.
Offshore QA- it is a great way to save money and a way to watch what is happening.

Metrics/ Reports
-User tells them what they want
-Getting metrics defined
-Getting top down buy in to help support
Everybody must be on the same page and agreements on all definition
-Common data is in warehouse
-What does metrics measure? Customer sales, profit, forecasts
-Be sure to segment users properly, what are they going to be doing with data, security may dictate # of marts
-Try not giving the user what they want instead give them what they need.

Process:
-Operational
-Put costs around batch windows
-SLA’s
-What level of data integration do you need?
-What is the time frame? How long do you keep data and what is necessary to keep.
-Make sure your company has defined the terms and code throughout the company, so every body is talking about he same thing.

Tools:
ETL- Ascential, DTS, Informatica
Reporting- Business Objects, MicroStrategy, BRIO, Cognos
Financial- Hyperion Enterprise
Relational tools- SQL, Oracle, Microsoft, DB2, Teradata

Comments/Benefits:
Experts / Good insight to many data warehousing areas
ETL Tools / Understanding the functionally – Options for quicker addition of new datamarts
Marketing our product / Needed more ideas on how to sale our product
Tools / What is available
Projects Pass & Fail / Info for next project
Data Mining / What else is available
What is needed to create data files / Gives an idea where to start
What causes databases to fail / Heads up on what to watch for
General overview of terms / Stressed importance of making sure everyone is speaking the same language
Strategies for making data warehousing successful / Job Security