Data Beliefs

To draw parallels with active investment management, adding value requires being at the same time contrarian and right. Therefore, starting in the new data governance and data quality role, I want to state somewhat contrarian "data beliefs" and then outline possible consequences for data governance.

These beliefs are about definitions of data and information, aboutness, classification and metaphors.

First, the terms data and information refer to the same thing for the practical purposes of managing it in an organisation. From one perspective, data and information are both signs that stand for something. Birger Hjørland, a fellow ISKO member, published a detailed analysis of definitions of data and information that are worth a look at.

Second, data and information have the property of aboutness. The "Orders" table in a database does not contain orders; it contains data about things people in an organisation call orders. A book by William Kent, first published in 1978, Data and Reality, is a great reference point here.

Third, ideas from faceted classification can be borrowed to state what data and information are about. You may be familiar with faceted classification by applying filters based on dimensions such as size, colour, or material in an online store. Still, the history of faceted classification goes back about 70 years or more. For a recent introduction to some of the faceted classification ideas, look at this talk by Leonard Will titled Creating a Structured Vocabulary.

Fourth, metaphors of substance and container are essential to understanding what is being managed and catalogued. When you hear people say, "let's fill this table with data", you hear one example of using metaphors - perceiving data as a substance and the database table as a container. Viewed this way, it is clear that we can practically only catalogue containers (or container types where containers are created and discarded often) and then state that data in these containers is about a particular concept. Metaphors We Live By by George Lakoff and Mark Johnson, first published in 1980, is an essential reference.

The consequence of these beliefs for setting up data governance is that you can start by using business language to build hierarchies of business concepts, and then use these business concepts to assign ownership and state in a data catalogue what data in various containers is about.