This article presents TUCANO (an acronym derived from the Portuguese for: Unified Techniques for the Categorization and Analysis of Online News), a hybrid methodological protocol for the analysis of online news corpora in Portuguese. Developed to address a challenge posed by non-probability sampling in Social Sciences, TUCANO seeks to delimit, organize, and explore sets of news articles without losing the interpretive depth of qualitative analysis. The protocol combines Natural Language Processing (NLP) techniques, such as tokenization, vectorization, and co-occurrence analysis, with a theoretical rationale oriented toward corpus construction and thematic subsampling. Its workflow is organized into six stages, ranging from the formulation of the guiding question to the definition of well-grounded subsets for qualitative analysis. Its applicability is demonstrated through a study on media coverage of the use of plastic straws and their impacts on marine life. With a lightweight, transparent, and accessible computational infrastructure, TUCANO enhances qualitative analysis and expands the critical use of computational tools in the Applied Social Sciences.
Date of publication
Sep 21, 2026
pt
Category:
Article
How can digital media be studied in times of information deluge? The TUCANO methodological protocol applied to documentary research on online news
by
Victor Felipe Mendanha de Lima
,
Frederico Marianetti Soriani