As the Keyboard Clicks...
The Collected Writings of Lori Roets Valone
On Track Rail News - August 22, 2008
From On Track Rail News, a Teradata in-house publication produced by Lori Roets

The Extraboard - An EDW Viewpoint - August 22, 2008
by Lori Roets

Garbage in…garbage out. We have all heard the old adage. Nowhere is it more true that in data warehousing. Good decision making is predicated on good information. When the integrity of data is compromised or even suspect, it makes any decisions based on that data suspect. As a result, data quality is a fundamental, but often overlooked, issue for any data warehouse. Railroads are no stranger to this issue. Born of the myriad of siloed legacy systems comprising the technical infrastructure of most railroads, “synonym” tables are the norm. These essential data translation guides bring together the 15 different ways to specific Timbuktoo or the umpteen  versions of customer name, Widgets R Us into a single data point for each. While the data is not cleansed at the source, at least translation – when applied  consistently – can provide a clean data stream within the EDW. The first step is truly understanding how big the data quality problem really is and resolving to find long-term solutions. Only by acknowledging they have a data quality problem, can railroads – and other companies – begin down the path to cleaner data and, ultimately, better decision making.