Coyuntura de la agricultura en Italia: metodología para un análisis textual en medios digitales

This study aims to present a replicable methodology based on natural language processing (NLP) techniques for the automated analysis of media discourse, and to apply this methodology to the case of the agricultural sector in Italy during the first half of 2024. A total of 164 news articles from seven Italian media outlets were analyzed to identify the central topics, predominant emotions, and discursive patterns. NLP tools in R were used to clean, tokenize, and vectorize the textual corpus, followed by the application of statistical models such as Term Frequency-Inverse Document Frequency (TF-IDF) and Latent Dirichlet Allocation (LDA) to identify topics, and the NRC Emotion Lexicon to classify emotions. Based on LDA, the thematic analysis identified two central axes in the media coverage: the institutional management of the Italian agri-food sector and the protests of farmers in the European Union. The first group focuses on sector development, highlighting themes such as technological innovation and public policies, while the second centers on farmers' protests and institutional responses. The sentiment analysis revealed a predominance of emotions such as trust (30%), anticipation (18%), and fear (13%). Peaks of negative emotional load occurred in February and March, associated with crises and protests, while the most positive news was concentrated at the end of April, coinciding with topics of innovation and international cooperation. The results significantly contribute to understanding how the media constructs the narrative about the agricultural sector. Additionally, the emotions that influence public perception are highlighted, which can be useful for finding policy references in the media related to the agricultural sector. This study highlights the potential of NLP for analyzing media discourse, providing a solid foundation for future research on the representation of agriculture in the media and its impact on public opinion.