ODArchive – creating an archive for structured data from open data portals
We present ODArchive, a large corpus of structured data collected from over 260 Open Data portals worldwide, alongside with curated, integrated metadata. Furthermore we enrich the harvested datasets by heuristic annotations using the type hierarchies in existing Knowledge Graphs.
We both (i) present the underlying distributed architecture to scale up regular harvesting and monitoring changes on these portals, and (ii) make the corpus available via different APIs.
Moreover, we (iii) analyse the characteristics of tabular data within the corpus. Our APIs can be used to regularly run such analyses or to reproduce experiments from the literature that have worked on static, not publicly available corpora.
T. Weber, J. Mitlöhner, S. Neumaier, A. Polleres, ODArchive – creating an archive for structured data from open data portals, in: Proceedings of the 19th International Semantic Web Conference (ISWC 2020), Lecture Notes in Computer Science Vol 12507, Virtual Conference (Athens, Greece), Springer (2020) 311–327