Social Media Data Mining with GeoAI: Opportunities and Challenges

Keywords: Social Media, GeoAI, Disaster Resilience, COVID-19, Responsible AI
Abstract Type: Paper Abstract

Abstract

Extracting valuable, fine-grained geographical information from social media data for various applications is challenging mainly because of technical difficulties in processing such big, biased, and noisy data. The breakthroughs in geographical artificial intelligence (GeoAI) provide solutions. Recent studies have demonstrated that newly developed GeoAI algorithms outperform traditional methods in several tasks relevant to social media data mining, e.g., place name recognition and text classification. Thus, analyzing social media data with novel GeoAI algorithms is promising in identifying location-based information precisely. This research elaborates on the opportunities and challenges of analyzing social media data with GeoAI. We investigated Twitter use as a case study to fulfill three objectives: (1) to develop a GeoAI-empowered framework for location-based social media analytics; (2) to demonstrate the applications of analyzing social media, e.g., disaster resilience and predicting COVID-19 outbreaks; (3) to pinpoint the challenges of using social media with GeoAI and propose solutions. This research will significantly advance the methods and applications of GeoAI and location-based social media analytics.