Themes and sentiments in conversations about food waste on Twitter: Proposal of a framework using neural topic modeling

Abstract

There is growing research interest in the role of social media in the food industry, and its potential has yet to be fully exploited. This study extracted and analyzed data from Twitter, now X, (2018 to 2023) using relevant keywords related to food waste (FW), identified frequently discussed topics, and determined the sentiments involved in tweets. With the aid of neural topic modeling techniques, this study found that the main topics discussed in this social network include “Food and Garden Waste Composting and Recycling,” “Global Food Waste and Management,” “Reducing Food Waste & Saving Money,” “Single-Use Plastic and Waste Management,” and “Creative Uses of Leftovers in Meals”. Some of the subject categories often discussed are FW Environmental concerns, FW events, Leftovers management, and FW household management. Most subject categories are associated with positive sentiments, which are associated with topics like tips and recommendations for reducing FW. Negative sentiments are related to discussions around the environmental and social impacts of FW. Our study found a gap in knowledge and awareness of food recycling practices. Our study extends previous research because the analyzed messages are not restricted to specific geographical regions. Moreover, our study also used a more varied dataset in terms of user profiles and employed a search string that comprised a more varied set of terms, making our search more thorough. A comprehensive framework of discussed topics and associated sentiments is proposed. Our study has crucial implications for the design of FW reduction campaigns and interventions in social media.