Data modelling methodology

The approach agreed to be adopted for MO.DWG conceptual modelling work is based around the INSPIRE methodology (D2.6), itself based on the results of the EuropeanRISE project. (Note the INSPIRE methodology document includes a meteorology example in Annex D.4.)

The methodology itself is quite straightforward, summarised in the diagram INSPIRE-methodology.png (taken from D2.6), and involves the following steps:

• Use case development: developing scenarios and use cases that define the scope of the targeted data harmonisation/interoperability
• Identification of user requirements/spatial object types: Identifying candidate information types/queries to satisfy the use cases to inform more formal model development
• As-is analysis: examination/evaluation of existing models and infrastructure as preparation for the next step
• Gap analysis: identifying gaps in existing models, undertaken together with the next step
• Data specification development: developing a formal application schema adopting the ISO TC211 framework, and based on the identified candidate spatial object types, existing models, gap analysis
• Implementaiton and testing:
• Cost-benefit analysis

In practice, the pragmatic experience of the INSPIRE Annex I data specs development process showed that some of these steps are more important than others (e.g. the gap analysis may not be that useful in its own right, but might be implicitly performed as part of the data spec development). A report is being prepared on the experience of applying it, and will inform the Annex II/III work (which includes meteorology/atmosheric conditions).

Identifying spatial object types

The second step is the identification of spatial object types - these are informatuion classes defined by the General Feature Model defined by ISO 19109. The ISO TC211 framework adopts this as the core of the data modelling framework. In essence it is an object-based modelling framework - 'spatial objects' aka 'feature types' are defined by the properties (attributes, associations, operations, generalisation). The output should be a 'first-cut' data specification.

One approach (discussed by the INSPIRE methodology) is the 'nouns method': "...consider the “nouns” representing real-world phenomena in user requirements and use case description as candidates for spatial object types. After eliminating duplicates and synonyms as well as nouns that describe concepts that are not spatial object types, the remaining list of nouns can be used as an initial list of spatial object types. By analysing the user requirements and the use case descriptions in more detail, the properties (i.e. attributes and relationships with other types) as well as the constraints will be specified for the individual spatial object types."

See also the Spatial Object Types derived from Use Cases.

-- JeremyTandy - 20 Apr 2011