Agricultural Research Using Social Media Data

Abstract

Core Ideas Social media data provide quantitative and qualitative data on agricultural practices.Twitter data accurately captures timing of crop planting across the United States.Maximum positive sentiment in planting tweets aligned with optimal planting period.Opportunities for mapping emerging agricultural issues and targeted intervention.Challenges include data availability and representativeness of social media users. The use of social media in scientific research is rapidly increasing, typically focusing on discrete events of interest to many people and/or spatially mapping a variable of interest. Relatively little research has been done on the utility of social media for monitoring the spatiotemporal patterns of day‐to‐day life, and none within the agricultural sciences. Here, I discuss the potential applications and limitations of social media data for agricultural research. As an example, I demonstrate the ability of Twitter to map state‐level corn and soy planting progress in the conterminous United States. Results compare favorably to traditional survey‐based crop progress monitoring, with mean absolute differences of <10% for most state‐crop combinations. I also highlight the additional contextual information available from social media data including factors contributing to replanting decision‐making and the evolution of farmer sentiment through time. Using analogs from other disciplines, I then discuss key opportunities and challenges for agricultural research using social media. Social media is particularly well‐suited for identifying emerging agricultural issues (e.g., weather, crop pests) and guiding extension and outreach directly to affected areas. However, limited data and unknown representativeness of social media users relative to the overall agricultural population are challenges which must be addressed for social media‐based agricultural research in the future.