Instagram Reactions to a Virtual Dining Companion:Qualitative Coding vs. Automated Sentiment Analysis

Abstract

Technologies designed to support solo eaters by providing them company are an emerging design space at the intersection of AI, affective computing, and everyday dining rituals. This study explores public reactions to such novel systems by analyzing Instagram comments on a viral post depicting a man dining with a virtual reality partner. Using a mixed-methods approach, we first applied automated sentiment tools - including TextBlob, VADER, NRCLex, and BERT - to quantify surface-level sentiment. However, these analyses failed to capture the layered, performative, and culturally nuanced nature of user responses. To address this, we conducted manual thematic coding of 719 comments, revealing a wide spectrum of emotional tones - from sarcasm and judgment to empathy, curiosity, and self-reflection. Our findings demonstrate that many seemingly "positive"comments were in fact mocking or ambivalent, and that public discourse about emerging technologies like virtual companionship is deeply shaped by generational expression, meme culture, and shifting social norms. Thus, social media can be a valuable source of opinions and preferences regarding novel interactive technologies - even speculative prototypes - but the collected data needs to be analyzed on multiple levels.