The following data story is defined in order to support the documentation of the development of the Kadaster Knowledge Graph (KKG) based on the IMX-Geo ontology. The ontology is published here and is visualised here. The following sections both highlight the content of the IMX-Geo KKG to date and as well as the relevant steps taken in its development.
SPARQL Construct Queries
The first step in the development of the IMX-Geo KKG is the transformation of instance data available using the models defined for various relevant key registers to the IMX-Geo model. This process is carried out by executing a number of SPARQL construct queries, defined based on mappings defined between the key registers and the IMX-Geo ontology. The complete list of necessary construct queries are defined here.
The following is an example of the SPARQL construct query defined to transform instance data mostly modelled using the BAG data model to the IMX-Geo Adres object type.
This sort of query is defined for all object types present in the IMX-Geo ontology. Some object types require that multiple construct queries be specified in order to ensure support for the performance of the construct query and to ensure that spatial relationships between the different key registers are properly captured during abstraction to a single IMX-Geo object type.
IMX-Geo KKG Content
IMX demo Zeewolde dataset
Links between BGT,- BAG- and BRT-elements are very handy in order to construct IMX data elements. The Kadaster Kowledge Graph (KKG) already contains these links making the conversion to IMX much easier. In this paragraph a sparql construct query has been developed that uses KKG data to create a subset of IMX focusing on the relations between Buildings, (main) addresses and Percelen.
IMX construct query using KKG
reusing KKG linksets to create IMX subset
The KKG has materialised geosparql sfWithIn relations between "gebouwen" and CBS Buurt information. As Buurt is part of a Wijk that is part of a Gemeente it is easy to constrain the query to extract only data from the Gemeente Zeewolde. Removing this restriction results in a dataset encompassing the whole of the Netherlands. This approach also enables to run this query per gemeente making it more performant.
Percelen en gebouwen
Data quality checks
The following map shows buildings that have and do not have connections to a Perceel. In red are the Gebouwen without Perceel.
