More About the Author
Fon Silvers began his career in Information Systems as a data entry clerk. Working his way through college a the University of South Florida (USF) he found a job with the Financial Management Association (FMA) entering data into a transaction system and membership database. When all the data had been entered, Fon's curiosity led him to look behind the data entry screens to understand the underlying application and database. For a student in Music Performance, this was all intriguing.
By the time he graduated with a Master's Degree in Music Performance, Fon knew two things. First, his degrees in Music Performance did not equip him to pay the bills. Second, he knew the underlying information system for the FMA. As a result, the FMA hired Fon to manage their information systems...and continue the data entry.
Working as a one-person Information Systems shop within the FMA, Fon learned about information systems from A to Z, including desktop hardware & software, networking, mainframe JCL & 4GL databases and report writing. After ten years with the FMA, and the advent of a family to support, it was time to move on. However, those degrees in Music Performance and ten years' experience were not quite enough to move to the next level.
Fon went back to USF to obtain an MBA in Information Systems. With that in hand Fon moved on to a Fortune 500 retail firm. Beginning as an entry level programmer, Fon's first assignment was as a COBOL programmer. This was the first of numerous platforms and languages Fon encountered.
After developing and supporting applications in the HR system for a few years, a data architect met with Fon over lunch and asked a simple question, "Would you be an ETL analyst for the Data Warehouse team?", to which Fon replied "Sure. What's an ETL? And, what's a Data Warehouse?".
Like all other challenges, Fon began with a pile of books. This time they were books written by Bill Inmon and Ralph Kimball. After a lot of research into Data Warehousing, the Data Warehouse team began building their first Data Warehouse. Successive efforts added subject areas and fixed gaps in existing subject areas. After two years, the Enterprise Data Warehouse had become a useful asset for the company. After a few years, the Enterprise Data Warehouse included real time data.
Fon leads the team of developers and analysts who support the Enterprise Data Warehouse as a production environment. Often included in development efforts, Fon provides the background to understand the existing Data Warehouse as well as the lessons learned along the way.
Writing those lessons into various articles led to Fon's book on Data Warehousing..."Building and Maintaining a Data Warehouse". Firmly rooted in the writings of Kimball and Inmon, Fon wrote "Building and Maintaining a Data Warehouse" for those who are where he was years ago, wondering "What is ETL? And, what is a Data Warehouse?", by providing the foundational principles, best practices and lessons learned the hard way.
Fon's newest publican, Data Warehouse Designs: Achieving ROI with Market Basket Analysis and Time Variance, presents three design solutions. The first design solution provides a Market Basket Analysis solution design. Market Basket Analysis provides the ability to provide and analyze the answer to a very simple question: "When a customers purchase a product, what else do customers tend to purchase?" The second design solution provides a Time Variant solution design. Time Variance, specifically Type 2 Time Variance as defined by Ralph Kimball, provides the ability to see transactions and events in a fact table in their original historic context. This design solution allows a data warehouse to simultaneously present transactions and events in their present context, what Ralph Kimball defined as Type 1 Time Variance. The third design solution combines the first two design solutions into a Market Basket Analysis application of Time Variant data. This gives the analyst the ability to consider the context at the original historic moment in time when performing Market Basket Analysis. Finally, the Market Basket Analysis of Time Variant data is able to a Type 2 Time Variant instance of an entity against the Type 1 Time Variant universe of all entities. It's a very powerful design. And, best of all, it is a design that is optimized to perform on any RDBMS capable of hosting a data warehouse. No huge additional capital procurement or expenditure beyond an existing data warehouse is required. If you have the architecture and infrastructure in place to operate a data warehouse, or are planning to obtain that architecture and infrastructure, then you have the required architecture and infrastructure necessary to implement a Market Basket Analysis application, Time Variant data, and a Market Basket Analysis application of Time Variant data.