Looking at the Data! What a Concept!
You cannot manage what you can’t see! This statement is especially true in Data Management, where an analyst cannot do a proper job without having some kind of read access to the entire range of enterprise data assets for which s/he is responsible.
Further, it is impossible to assess data quality without looking at the actual production data, since quality is what happens to data as a result of business activity (and I.T. manipulation).
Data management operates in a radically different paradigm compared to what the majority of I.T. people (and personalities) use. By their nature, hardware enthusiasts and programmers have little interest in the data of the enterprise. Yes, they may test a record or two coming out of some process, but that’s about it. They are less interested in whether or not the data reflects business reality; they want to move on to other creative coding activities.
Case in point: As the DQ manager on a project, I once asked for the ability to execute queries against a database that was being created and loaded. The programming manager frowned. “No, just tell us what reports you want, and we will write the code to produce them.”
Wrong! Nothing can be further from what I seek in this article. As data professionals, we need to have unfettered, read-only access to the data. I don’t want the ability to change any piece of production data. I don’t want anyone blaming me if it changes mysteriously. But I do want to see what is going on in there.
Data Quality and Data Management people are interested in data, and need to be able to see it.
Michael Scofield provides practical tips and guidelines for analysts tasked with data profiling responsibilities.