TopoText 2.0, the second iteration of TopoText, is a prototype that performs “text to map” conversion; users can simply upload a plain text and visualize it as a map, as well as collocate places and words in order to trace most frequent topics around specific places across texts. TopoText generates a comma-separated values (CSV) file of the geographic entities included in each text matched with lat-long coordinates, which can then be exported, allowing reusability in other environments. It also allows users to correct mismatched geographical points through the “human-in-the-loop” approach where the user can access the generated dataset and select alternative locations with the same place name. They can also annotate geographical points and visualize these annotations on the map interface. In this sense, TopoText 2.0 is both a mapping tool and an open geodata provider.
Randa El Khatib (project lead; University of Toronto), Julia El Zini (developer; American University of Beirut), David Joseph Wrisley (NYU Abu Dhabi), Mohamad Jaber (Google), and Shady Elbassuoni ( American University of Beirut).
Keywords: open geodata generator, “human in the loop”, geoparsing, experimental prototyping, digital mapping