Mapping Antarctica through text and location-based social media analysis

Antarctica is a continent where people do not actively live due to its location and climate characteristics. Consequently, geographic information production in Antarctica is limited from various perspectives. Conventional mapping technologies, such as remote sensing, photogrammetry, and ground surveying methods, contribute to the digital database of Antarctica. In addition to these techniques, new mapping technologies have emerged in the last decade. One of these technologies, defined as "crowdsourced mapping," considers humans as sensors perceiving their environment. Social media platforms contribute to this new mapping technology through the data they provide, and the information inference algorithms generated. However, data produced in social media sources contain uncertainties in terms of data quantity, completeness, and bias compared to regularly generated data in traditional technologies. The unstructured data and uncertainties in data quality of social media lead to changes in information extraction algorithms, and the significance of the generated results is also debated. In this study, social media data collected for the Antarctica continent are discussed in terms of the text-based information extraction algorithm and its results. In this study, a text and location- based information extraction algorithm is presented for the Antarctica continent for the first time. In addition, the social media data analysis methods suggested in this study are repeatable for extracting geographic names and mapping current topics in regions with data scarcity and mixed natural language use.