Best Practices

Example Cubes

For published examples see https://lindas.admin.ch/datasets/ with the tag "Cube".

Naming Conventions

• For Target Property (in CSV Mapper)
  • lowerCamelCase (see here).
  • Use english terms if possible.

Data with municipalities, cantons and countries

• use three different shared dimensions (Municipalities, Cantons and Countries) on the concept that contains the values for municipalities, cantons and countries (by using the + sign)
• use schema:alternateName as Identifier property if your data uses abbreviations like ("BE", "ZH", "CH")
• choose as Hierarchy Switzerland - Canton - Municipality and leave all the other fields
• retransform the cube

Referencing read-only Shared Dimensions in other Shared Dimensions

When linking terms in a Shared Dimensions the usual approach would be to describe the relationship of the parent and child bottom up. Sometimes these relationships will also inclube terms from other Shared Dimensions (Connecting Shared Dimensions. Meaning that for example the canton of Bern (subject) would be containedInPlace (predicate) Switzerland (object). However if the terms of the subject in this case canton is defined as a read-only Shared Dimension. This would not be possible. Describing the relationship via a top-down approach is a work around for this problem. Meaning we would describe Switzerland instead of Bern in our previous example and the predicate would change to containsPlace to match the description.