Meeting notes - November 16, 2022

[InKyungChoi]

Participants: Francine Kalonji, Flavio Rizzolo, Zoltan Vereczkei, Andrea Petres, Ayman Hathoot, Florian Vucko, Krishnan Ambady, Jenny Linnerud, Juan Munoz, Ayman Hathoot, Cecilia Casagrande, Wai Kit (Jackie) Si Tou, Christopher Jones and InKyung Choi

Issue #6 Missing GSIM class – quality indicators: Discussion on the mapping

• Referential Metadata (RM) Subject Item is at the instance level, and it can be related to more than just one RM Attributes (e.g., in this example of table, footnote 1 and 2 are related to the data table, while footnote 3 and 4 are related to geography). Note that RM Content Item would relate to only one RM Subject Item.
• In Data Structure, Data Structure Component (e.g., measure, identify) is defined through Represented Variable. This can happen if we have, e.g., table-like referential metadata. Here, RM Attribute is not just referencing a Represented Variable, it is defined by Represented Variable.
• Some RM Attributes will point to Represented Variable, but not all so it should be optional. Maybe we can use “Represented Variable can play a role of RM Attribute”.
• RM Subject should point to a GSIM classes (abstract general ones) while RM Subject Item should point to an instance version of that class. For variable, RM Subject can be a Represented Variable, RM Subject Item can be actual Represented Variable (“total residents”; but not Instance Variable).

Issue #28 GSIM Structure Group definition / explanatory text update

• Current distinction between Dimensional Data Point (DP) and Unit DP is rather forced. The fact that they are both container of value is the same, only different is where this container is put (in Unit Data Set or in Dimensional Data Set). This distinction is not necessary.
• Action: to remove Dimensional Data Point and Unit Data Point. Link Data Point with Unit and Population with “is observation of” but with cardinality 0,1 (e.g., you can have a Unit Data Set but without way to link to any Unit because the connection is removed due to confidentiality reason)
• Does it make sense to differentiate Unit Data Set and Dimensional Data Set? Maybe we can also remove them and make an attribute in Data Set to indicate whether it is dimensional or unit.