Authors: Molly Miranker*, Texas State University - San Marcos, Alberto Giordano, Texas State University San Marcos
Topics: Human Rights, Geographic Information Science and Systems
Keywords: Humanitarian GIS, Corpus linguistics, Text mining
Session Type: Virtual Paper
Start / End Time: 11:10 AM / 12:25 PM
Room: Virtual 32
Presentation Link: Open in New Window
Presentation File: Download
The methods and tools of Geographic Information Sciences (GISci) are increasingly used in forensic humanitarian projects. Using corpus linguistic(CL)/natural language processing (NLP) and Qualitative Spatial Representation (QSR)/Semantic Triples (ST) methods, social and media releases from US Border Patrol (CBP) were examined to gain an understanding of the death of undocumented border crossers (UBC) along the US-Mexico border.
Migrant death at the border has been called a humanitarian crisis due to the amount of death and exhaustion of resources. A persistent obstacle to responding to UBC death is inconsistent case documentation (Spradley et al. 2019). CBP Twitter and media releases were analyzed as potential sources to link UBC case information.
Results indicated that CL/NLP and QSR/ST have the potential to increase and improve UBC case identification by providing a framework for searching for key terms or themes throughout multiple textual sources. CL/NLP showed that CBP social media focused on drug confiscation and were of limited use for tabulating incidences of UBC deaths. QSR/ST visualizations showed which CBP Stations most frequently reported UBC recoveries and with whom they collaborated.
These methods are part of the methodological toolkit needed to create what we call “Humanitarian GIS,” which we define as the application of spatial analytical perspectives and tools in various human rights topics and events such as genocide studies and spatial forensics. Within this toolkit, CL/NLP and QSR/ST highlight spatial relationships that are not necessarily mappable in a traditional GIS setting and allow detection of patterns across large corpora of heterogeneous information.