Social media posts as a source of ecological information over time: using Twitter (X) as a proof of principle

Abstract

Acquiring data about ecology over long periods of time and large geographical areas is often difficult, expensive, and takes a long time. Here, we explore whether, through millions of daily posts on social networks, we could be constantly and unconsciously accumulating data regarding important biological patterns over the years. Data analysis from generalist social networks has been successfully applied in different research fields, most notably in political science and epidemiology. Here, we evaluate their potential and drawbacks for studying biological patterns of other organisms over time. To that end, we used millions of posts on Twitter (currently X), over 11 years, on four different insect groups: cockroaches, crickets, Monarch butterflies, and mosquitoes. Our results show that through millions of daily tweets, we could identify temporal periodic patterns that reflect their unique known phenology of the four insect taxa studied. Given the 11-year span of the dataset, we could also track changes in their patterns over the years that might be related to environmental factors. Using a sentiment analysis, we could also characterize the emotions of people towards these animals, which is important for the design of awareness campaigns. We also discuss the limitations of these data, such as the potential that social media data might have for spatial tracking of migrations. In summary, we show compelling evidence for the use of social network data for biologists and provide a theoretical and practical framework for exploring these sources of data.