Hierarchy in Reference Data Tables

Sometimes there can be hierarchies in reference data tables.  One example is the US State Department Travel Advisories web page.  The Travel Advisories table has a 1 to 4 scale of Travel Advisory Levels, where 1 is the least severe and 4 is the most severe.

This is an example of hierarchy in reference data tables.  The Travel Advisory Levels table is used to categorize each individual Travel Advisory record.  However, there are different types of hierarchy in reference data tables.  The most common appears to be the traditional generalization hierarchy.  This is a very ancient pattern, dating back (in terms of visual representation) to the Tree of Porphyry, devised in the 3rd Century.  In it, the higher levels are more general concepts that cover the more specific concepts in the lower levels.  This process of logical division continues until the species level is reached, at which point only individuals (i.e. instances) occur at a lower level.

This pattern is seen in Sector – Subsector relationships in industrial classifications, like NAICS.  The Travel Advisory Levels, however, is different.  It is a categorization that is independent of the attributes possessed by what it is categorizing (i.e. countries).  We might think of some way of further categorizing the Travel Advisory Levels themselves, e.g. into those we take seriously and those we do not.  This would be an even higher level of independent categorization. 

This brief examination of hierarchy in reference data shows that there are diffident patterns of hierarchies in reference data.  Although this area requires a lot more theoretical work, it is very likely that these different types of hierarchy are going to have different governance needs, and this is going to have practical impact for reference data management.