EML's concept of measurement types and the Steven's typology is not always familiar. A simple flow chart could help explain the choices and give examples, such as:
flowchart TD
A(Variable X) --> B{Values naturally ordered?}
B --> |Ordered| C{"Equal magnitudes between values?
(6 - 5 == 11 - 10)"}
C -->|Equal - Quantities| D{Represents date or time?}
B -->|Unordered| F("fa:fa-car Nominal
Left / Right")
C -->|Unequal| G("fa:fa-car Ordinal
Low / Medium / High")
D -->|Yes| H("fa:fa-car DateTime
2023-05-13T14:32:11Z")
D -->|No| I{Has meaningful zero?}
I -->|Not Meaningful| J("fa:fa-car Interval
40 C")
I -->|Meaningful| K("fa:fa-car Ratio
12.2 m")
Feedback welcomed on this representation, and better ways of explaining each. In EML we mainly use this to differentiate what kind of attribute metadata is needed to describe the variable, so the interval/ratio distinction isn't really needed. But otherwise, for the rest of them they result in different metadata needs. Should we add something like this to the documentation?
EML's concept of measurement types and the Steven's typology is not always familiar. A simple flow chart could help explain the choices and give examples, such as:
Feedback welcomed on this representation, and better ways of explaining each. In EML we mainly use this to differentiate what kind of attribute metadata is needed to describe the variable, so the interval/ratio distinction isn't really needed. But otherwise, for the rest of them they result in different metadata needs. Should we add something like this to the documentation?