Sentiment Analysis of Human-Arthropod Interactions

Abstract

Humans' intense fear of spiders is well-known. There are studies that show it exists and theories on why. However, the body of work on how warranted this fear is based on potential harm by spiders is much smaller. Few species pose real danger to humans relative to all worldwide and this handful is not globally distributed. Additionally, due to lacking funding/researchers, there is little knowledge on ranges and distributions of spiders compared to locations of people that fear them. The proliferation of social media provides new opportunities for mapping and opinion data collection. Methods were synthesized to collect and analyze point data geolocated inside the United States from Twitter (Chapter 2) and iNaturalist (Chapter 3). Literature was reviewed to find which spiders are medically-important: those generally agreed upon by scientists to have venom likely to harm humans. Data were normalized to account for spatial userbase differences. Most tweets about spiders in the US were posted from major cities during peak spider activity in late spring/early fall. However, geographic locations of these online activities and sentiment of relevant tweets had little overlap with ranges and distributions of spiders that could actually harm them. This implies that many people discuss spiders, often negatively, regardless of their likelihood of encountering any of medical importance. Prior to this research, there has not been a spatially-explicit arachnophobia test. This methodological syntheses constitutes a new paradigm for creating estimate species maps and comparison to opinion data, expanding public outreach and conservation possibilities.