Enabling Spatio-temporal search in Open Data

This is a project by the Institute for Information Business at WU Vienna (Vienna University of Economics and Business). The aim of this work is to annotate a large corpus of tabular datasets from open data portals, and enable structured, spatio-temporal search over Open Data catalogs through a spatio-temporal knowledge graph, both via a search interface as well as via a SPARQL endpoint. This knowledge graph is constructed by integrating and interlinking

  • the openly available geospatial datarepository GeoNames
  • the map service OpenStreetMap
  • country-specific sets of postal codes
  • the European Union’s classification system NUTS
  • Periodo, a repository of historical, art-historical, and archaeological periods
  • the geo-cooridinates and temporal information available in Wikidata
This base knowledge graph is used to add semantic labels to the open datasets, i.e., we heuristically disambiguate the geo-entities in CSV columns using the context of the labels and the hierarchical graph structure of our base knowledge graph.

Currently, this showcase user interface contains CSV tables from several European data portals indexed at the Open Data Portal Watch, e.g.:

This is a research project and we try to add more features and fix bugs. Any feedback, bug reports, and other comments are very welcome.

Related Publications

Enabling Spatio-Temporal Search in Open Data
Sebastian Neumaier and Axel Polleres
Geo-Semantic Labelling of Open Data
Sebastian Neumaier, Vadim Savenkov, and Axel Polleres

In 14th International Conference on Semantic Systems (SEMANTiCS), Vienna, Austria, September 2018. to appear.

reboting.com: Towards geo-search and visualization of Austrian Open Data
Erich Heil and Sebastian Neumaier

In Proceedings of the 15th European Semantic Web Conference (ESWC2018 – Posters and Demos Track), Heraklion, Crete, Greece, June 2018.

Source Code

The source code is available on

Data Dumps

The data contained in the SPARQL endpoint is also available for download as n-quads: