Stage 1. Automated inclusion of datasets

The following tasks have been performed to automatically analyse the coverage of the SSN ontology in datasets:

1. To select the datasets to be included in the study. The datasets selected are those that are included in the following registries:
   • LOD Laundromat: It includes datasets collected by catalog-specific scripts (e.g., the Datahub ones) and by hand.
   • LOD Cloud Cache: It includes crawled data from various RDF and Linked Data resources from around the Web, as well as data extracted using Virtuoso Sponger from web pages.
   • LODStats: It includes the datasets from the data.gov, publicdata.eu and datahub.io data catalogs.
2. To identify which vocabulary terms from the SSN ontology are used in each dataset. In this task the Loupe system has been used to analyze the vocabulary usage patterns stored in LOD Laundromat (657,902 datasets). Besides, LOD Cloud Cache (3,203,793 documents) and LODStats (2,005 datasets) have been analysed using SPARQL queries. It must be noted that the identification has been based on the direct mention of the terms in the datasets; it does not include the cases in which some SSN term is specialized in another ontology and in the data the specialized term appears and not the SSN one.

Stage 2. Manual inclusion of datasets

The automated approach is not enough for analysing the usage of the SSN data, since these data may not be openly published in the Web, may be part of streams of data that are not persisted, may be used internally in data processing infrastructures, etc. Besides, the different vocabulary terms could have been extended and the datasets could contain the extended terms and not the ones from the SSN ontology.

Therefore, in a second stage an open call has been issued for people to submit their own datasets to be taken into account.

If the content of the datasets can be retrieved, the previous analysis can be performed over them. If not, people willing to submit a dataset had to state which terms of the SSN ontology are used in their datasets.

Analysis of usage in datasets (Consumers)

The following table presents the datasets that use the SSN ontology that have been found, along with their URIs.

The following table presents the coverage of the modules of the SSN ontology in the different datasets.

The following table presents the coverage of the vocabulary terms of the SSN ontology in the different datasets (Consumer Implementations).

Stage 1. Automated inclusion of ontologies

The following tasks have been performed to automatically analyse the coverage of the SSN ontology in ontologies:

1. To select the ontologies to be included in the study. The ontologies selected are those that reuse the SSN ontology, that are included in the LOV ontology catalogue, and that are available in the Web.
2. To identify which vocabulary terms from the SSN ontology are used in each ontology. This task has been performed manually.

Stage 2. Manual inclusion of ontologies

The automated approach is not enough for analysing the reuse of the SSN ontology in other ontologies, since these ontologies may not be published in the LOV catalogue.

Therefore, in a second stage an open call has been issued for people to submit their own ontologies to be taken into account.

If the content of the ontologies can be retrieved, the previous analysis can be performed over them. If not, people willing to submit a ontology had to state which terms of the SSN ontology are used in their ontology.

Analysis of the usage in ontologies

The following table presents the ontologies that reuse the SSN ontology that have been found, along with their URIs.

The following table presents the coverage of the modules of the SSN ontology in the different ontologies. The second row indicates whether the ontology imports the whole SSN ontology or not.

The following table presents the coverage of the vocabulary terms of the SSN ontology in the different datasets. The second row indicates whether the ontology imports the whole SSN ontology or not.