A methodological approach for soundscape attributes translation using Twitter data: A case study in the Portuguese language

ISO 12913-1:2014 proposed a standardized method for approaching soundscape research and formulated eight Perceived Affective Qualities (PAQs) in the English language to assess the soundscape. However, studies conducted in non-English speaking countries have revealed socio-cultural disparities in soundscape evaluations, which may be influenced by how certain words and phrases are used and understood. This paper endeavors to establish a method for validating the PAQs in languages other than English by leveraging their frequency in social media, using the Portuguese language spoken in Brazil as a study case. In the initial phase of the method, English-to-Portuguese translation experts determined PAQs' synonyms. Secondly, the method entailed searching for these synonyms on the Twitter platform and employing data mining techniques to filter and extract contextual meaning. As a result, in most cases, the validation process aligned with the choices made by the translation experts, except for the Vibrant and Uneventful PAQs, which exhibited alternative options more frequently cited by social media users. Eventful and Uneventful PAQs elicited a diverse array of synonyms from experts and garnered fewer tweets compared to other descriptors, suggesting that these synonyms may not adequately capture soundscape perception in Brazilian Portuguese. Supported by this evidence, we propose a revised arrangement of six PAQs, retaining the two-dimensional model from ISO 12913-3:2019 while eliminating the Uneventful-Eventful axis. This restructuring holds promise for providing a more accurate representation of soundscape perception among native Brazilian Portuguese speakers. Future directions may include refining the data mining technique and applying the methods to other Portuguese-speaking regions, ultimately enhancing the understanding of soundscape assessment in diverse linguistic and cultural contexts.