This paper demonstrates LODChain, which is an online service for tackling the problem of low connectivity for a given RDF dataset, by strengthening its connections to the rest of LOD Cloud. The demo focuses on presenting the connectivity analytics, visualizations and enrichment services for the user/publisher, which are ofered for several parts of the dataset, e.g., for owl:sameAs mappings, entities, schema elements and triples. Moreover, we show how these analytics can be exploited for improving the connectivity of a dataset, and thus its discoverability and reusability, by using a scenario with a real RDF dataset.
Current approaches for dataset search and discovery are mainly metadata-based and ignore
the content of the datasets [
This demo paper is a complement to an accepted paper of ISWC’22 Resource Track [
© 2022 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
The rest of this demo paper is organized as follows: §2 introduces the related work, §3 describes in brief the process of L O D C h a i n , §4 presents all its services, and §5 concludes the paper.
There are several approaches for aiding the interlinking, discoverability and reusability of RDF
datasets. Concerning interlinking, applications like WIMU [
We present the ofered services, divided in six categories by using a real dataset from publications
domain, i.e., WW1LOD [
Category 1. Analytics over owl:sameAs Mappings. The objective is to evaluate for the given dataset if L O D C h a i n a) inferred new mappings through the computation of owl:sameAs closure, and b) detected possible owl:sameAs errors, i.e., L O D C h a i n checks if an entity is equivalent with two or more real entities in LODsyndesis. L O D C h a i n ofers a chart analyzing the results of the computation of closure, including the number of a) possible errors, and b) owl:sameAs relationships before and after the computation of closure. In case of detecting errors, they can be downloaded from the user for aiding the process of fixing them. As Fig. 2(a) shows, we inferred thousands of owl:sameAs mappings for WW1LOD, without detecting errors, which is very positive for its connectivity, i.e., we expect that more connections have been discovered.
Category 2. Analytics over Entities. The objective is to evaluate how connected is the given dataset to the rest of LOD Cloud, and to quantify the gain of the computation of owl:sameAs closure. L O D C h a i n provides charts for the number of unique/common entities, the connections before and after the closure, the top-10 connected datasets, and its connectivity compared to datasets of several domains. For instance, Fig. 2(c-d) shows that WW1LOD shares 825 entities with other datasets (such as DBpedia and Wikidata), whereas 25 new inferred connections were discovered through L O D C h a i n . Moreover, a graph visualizes all the connections of the dataset (Fig. 2(f)), where the red nodes depict the old connections and the green nodes the new ones (due to closure), and the labels of edges the number of common entities. From these charts, we can see that the connectivity of WW1LOD was highly increased. On the contrary, in a scenario where a dataset has either zero or few connections, even after the computation of closure, it could be an indication of bad/low connectivity, e.g., few or/and incorrect links to other datasets. Finally, charts like Fig. 2(e) can be used for quantifying the connectivity of a dataset comparing to other domains, e.g., WW1LOD is connected with datasets from 5 diferent domains.
Category 3. Analytics over Schema Elements. The goal is to evaluate if existing
ontologies are reused, since it is important for a) comparing the values of the same fact for an
entity [
Category 4. Analytics over Triples. The objective is to evaluate whether a dataset shares common triples (or facts) with others, and if there are available complementary facts for its entities. L O D C h a i n ofers charts showing the number of common/unique facts, and the top-10 datasets ofering the most common and complementary facts for the entities of the given dataset. Fig. 2(i) shows that WW1LOD contains 368 common facts with at least one RDF dataset. Certainly, it is not required for a dataset to contain common facts with other ones for being included in the LOD Cloud, however, they can be of primary importance for data verification. On the contrary, the existence of complementary facts can be useful for enriching the content of the given dataset, e.g., we found 362,339 complementary facts for the entities of WW1LOD.
Category 5. Dataset Discovery and Selection. In many cases, the publishers desire to
select and integrate their dataset with only K datasets, i.e., since it can be time-consuming to
export and to integrate with any available relevant dataset [
Category 6. Data Enrichment. L O D C h a i n gives the option to the publisher to export the results of the connectivity analytics and to enrich the given dataset with additional data for creating an enriched version of their dataset. In particular, one can export in RDF format all the common entities, schema elements and facts (and their provenance), the inferred owl:sameAs relationships, complementary facts, and rich metadata including the results of the analytics. All these data (or any subset of them) can be used from a publisher for improving his/her dataset.
We presented the connectivity analytics, visualizations and enrichment services, ofered by L O D C h a i n ; a web application for strengthening the connectivity of an RDF dataset over hundreds of datasets. As a future work, we plan to extend L O D C h a i n for ofering more services, e.g., finding connections for almost or totally disconnected datasets (by performing instance matching), and services for further aiding the process of fixing connectivity errors, e.g., owl:sameAs errors.
This work has received funding from the European Union’s Horizon 2020 coordination and support action 4CH (Grant agreement No 101004468).