The current media landscape continuously exposes us to a large amount of information. However, human attention capacity is limited, making the study of collective attention dynamics particularly relevant. This Master’s Thesis explores the characterization, through a filter, of country-specific interest patterns based on how Spain, Brazil, the United States, and Germany respond to news of different types. To this end, we propose a linear model that relates the thematic composition of news items, obtained using a classifier, to the attention they generate, measured through Google Trends using representative queries and a calibration procedure. The results show that differentiated filters can be recovered across countries, with both common structures and differences in the relative importance of specific topics. Although the estimated filters show high robustness to different definitions of the response, these patterns should be interpreted within the analyzed time period, and broader temporal coverage is needed to determine their persistence. Overall, this work highlights the potential of the proposed approach to quantitatively study collective attention and explore differences across countries in their response to different information stimuli.