摘要: Many datasets nowadays contain links between geographic locations and natural language texts. These can be geotags, such as geotagged tweets or Wikipedia pages, in which location coordinates are explicitly attached to also place mentions, those news articles, travel blogs, historical archives, texts implicitly connected the mentioned places. This kind of data is referred geo-text data. The availability large amounts brings both challenges opportunities. On one hand, it challenging automatically process this due unstructured complex spatial footprints some other offers unique research opportunities through rich information contained special geography. As a result, facilitates various studies especially data-driven geospatial semantics. paper discusses related concepts. With focus on research, systematically reviews number that have discovered multiple types knowledge from Based literature review, generalized workflow extracted key for future work discussed.