Graphlet Column Hierarchies
Hierarchies are ubiquitous because they help us to organize and understand what happens around us. They may be ‘physical’ (as in the case of dimensions relating to ‘space’ and ‘time’) or more ‘conceptual’. An example of a spatial hierarchy is ‘Continent > Country > Region’, and a temporal hierarchy might be ‘Decade > Year > Quarter > Day’.
Many JMP customers make tangible products, and measurement hierarchies that reflect the structure of their products are the rule rather than the exception:
- Design revision > Batch > Lot > Wafer > Die
- Product family > Campaign > Batch > Sub-batch
- Factory > Material Supplier > Machine > Run > Within-Run
- Machine Type > Machine ID > Batch ID > Mould Position
Although often ‘obvious’, note that such hierarchies are often not absolute since they are conditioned by what is actually measured and what the purpose of any analysis or reporting might be. So, the ability for users to quickly establish and reform hierarchies from variables is essential. JMP’s Graphlets provide valuable capabilities for end users, often with no coding, but people probably still need to learn how they can be used to best effect. Although the ‘Next in Hierarchy’ column property provides the means for a determined user to build a hierarchy, this feature is a little hidden, takes too many mouse clicks, and does not facilitate the effective use of Graphlets even if someone is already convinced of their value.
Further development of the ‘Define Hierarchy’ add-in would streamline the use of Graphlets in the usage patterns mentioned above. It goes beyond a convenience feature, and (with appropriate promotion) would support their more effective use and our wider goal of helping users to raise the effectiveness of the way they are able to work with data.
Unlike the attached add-in, any development of it should support hierarchies across virtually joined tables and provide the facility to properly manage all the column properties that it defines across all tables.
