Geospatial Knowledge Graphs

and Reasoning, Natural Language Processing, Machine Learning, and others. In addition to this historical survey, interesting recent surveys on knowledge graphs are Hogan et al. [2020] and Weikum et al. [2021]. The idea of large ontologies and KGs goes back to the seminal projects CYC [Lenat et al. 1990] and WordNet [Miller 1995]. The first modern KGs appeared around 2007 with the development of DBpedia [Auer et al. 2007], YAGO [Suchanek et al. 2007] and Freebase [Bollacker et al. 2008]. Google bought Freebase in 2010 and used it to build the Google KG, which today powers its search engine. As a result, when we ask Google “What is the height of Mount Olympus” today, we get the pre­ cise answer 2,917 meters, instead of links to Web pages where the answer could be found as we would get in the past. But KGs do not just power today’s search engines; they play an important role in many other large industries [Noy et al. 2019, Sequeda and Lassila 2021]. In this chapter, we are interested in knowledge graphs that contain rich geospa­ tial knowledge that can be used to answer geospatial questions such as “Which river crosses the city of Larissa?” or “What countries border Greece to the north?” or “What is the distance between Athens and Salonika?” The next chapter dis­ cusses the problem of answering geospatial questions such as these over geospatial knowledge graphs. Geospatial Knowledge Graphs