WikiEvents Dataset from January 2020 to December 2022

WikiEvents is a knowledge graph based dataset for NLP and event-related machine learning tasks.

This dataset includes RDF data in JSON-LD about events between January 2020 and December 2022. It was extracted from the Wikipedia Current events portal, Wikidata, OpenStreetMaps Nominatim and Falcon 2.0. The extractor is available on GitHub under semantic-systems/current-events-to-kg.

The RDF data for each month is split onto four graph modules each:

The <strong>base</strong> graph module contains events, event summaries with references from named entities to Wikipedia articles.
The <strong>ohg</strong> graph module with all one-hop graphs (ohg) around the referencend Wikidata entities.
The <strong>osm</strong> graph module which contains spartial data from OpenStreetMap (OSM).
The <strong>raw</strong> graph module containing the raw HTML objects of events and article infoboxes.

This repository additionally includes two JSON files with training samples used for entity linking and event-related location extraction. They were created using queries to the WikiEvents dataset uploaded into this repository.