Sentiment analysis in rural areas: a systematic review of applications and approaches

Abstract

Rural depopulation and territorial imbalances are ongoing challenges that require innovative tools for understanding community needs and informing evidence-based policy. Sentiment Analysis (SA) has emerged as a promising methodology for extracting emotional information from large volumes of digital text data. However, despite its growing methodological relevance across the social sciences, its application to rural contexts is limited, and no prior systematic review has examined the application of SA within rural research. This article aims to assess the usefulness and potential of SA in rural research through a systematic review of existing applications and a critical analysis of its methodological contributions. The systematic review was conducted in accordance with PRISMA guidelines. Eligible studies applied SA to textual or social media data and focused on rural areas or explicitly compared rural and urban contexts. Searches were performed in Scopus and Web of Science (WOS) up to March 2025. A total of 47 studies were included and analysed. Data extraction was performed from the reviewed studies, collecting information on the location, study type, authors’ discipline, and methodological quality using the Newcastle-Ottawa Scale (NOS) and a consensus review by the research team. Key results were extracted, including areas of application of SA, methodologies employed, and data sources. Additionally, a thematic analysis was conducted considering the main areas of application of SA and its limitations. The studies reveal a developing field where SA is increasingly applied across various domains, particularly in rural tourism and public health, as well as in agriculture, food policy and social well-being. A significant geographic concentration of research is observed in China and the United States. The studies employ various SA processing methodologies (lexicon-based, machine learning, deep learning and hybrid approaches) and diverse data sources (social media, surveys, digital reviews). Methodological quality was generally high, with 85% of the studies scoring 7 or higher on the Newcastle-Ottawa Scale. Furthermore, the analysis identified notable differences in perceptions between rural and urban areas, emphasizing the importance of considering local contexts in public policy discussions. These findings contribute to a deeper understanding of the role of SA in rural development and offer valuable insights for future research and policy formulation aimed at addressing the specific needs of rural communities.