Eating disorders are a global health concern as they manifest in increasing numbers across all sections of society. Social network platforms have emerged as a dependable source of information about the disease, its effect, and its prevalence among different sections. This work lays the foundation for large-scale analysis of social media data using large language models (LLMs). We show that using LLMs can drastically reduce the time and resource requirements for garnering insights from large data repositories. With respect to ED, this work focuses on understanding its psychological impacts on both patients and those who live in their proximity. Social scientists can utilize the proposed approach to design more focused studies with better representative groups.