The visibility of states in news media is relevant for a broad range of research agendas in International Relations, IPE, Political Science, and beyond. This dataset – coming from the project [Global Flows of Political Information (GLOWIN)][1] – is the first to offer comprehensive global media visibility scores for states representing more than 97% of the world population and 99% of the world's GDP. Unlike previous research, which often concentrated on the analysis of news media content in mostly Western states, the present global visibility scores equally reflect the visibility of individual states in news media across the vast majority of the states of the world – the same 147 states that are analyzed. ***Data*** --------------------- To estimate the states’ visibility, we analyze the textual content of a carefully selected sample of online news media from the 147 states, in total more than 3.4 million unique news articles published between January 2018 and December 2021, a total of close to 12.6 million data points (an article/media outlet can be read in multiple states). The number of articles analyzed per state varies between less than 10,000 for the fifteen least represented audience states to more than 200,000 for the fifteen most represented audience states. English language content accounts for close to 700,000 news articles (20%), the remaining more than 2.7 million (80%) was machine-translated using Google Translate. A lexicon with typically two search terms per state was used – one for the common English state name and one for the respective adjective (e.g. Australia, Australian, not Commonwealth of Australia). Multiple or alternative names were accounted for (e.g. Czech Republic and Czechia). In specific cases (e.g. Georgia, Dominica), supervised machine learning and generative AI tools were used to disambiguate the references. The keyword detection was case-insensitive, regular expressions were generally used, especially to avoid false positives (e.g. Mali, Somalia, malicious). For the USA, the following search terms were used (with regular expressions): United States of America, USA, United States, the US, U.S., American. Uses of the adjective American beyond references to the USA (e.g. South American) were discarded. The aggregate global visibility scores represent unweighted arithmetic means of individual states’ media visibility scores across each of the other 146 audience states in the dataset. These underlying visibility scores were calculated, for each individual audience state, as the share of articles, in the given audience state across 2018-2021, that contain at least one reference to the foreign state analyzed (the visibility of which is being estimated). The calculations do not reflect the readership size of the different media outlets, though only outlets ranking nationally among the top 500 most widely accessed web domains, based on Amazon Alexa Ranking (data from 2021-2022), were used. ***Variables*** --------------- Variables included in the dataset are: **iso3**: States are identified by their iso alpha-3 code. **VisibilityRank**: Ranking of states by variable Visibility, from the highest to the lowest. **Visibility**: The variable represents the percentage [0-100] of articles in which at least one reference to the state appears, calculated as the unweighted arithmetic mean of individual audience states’ scores. For each audience state, the number of all articles containing at least one reference to the analyzed state is divided by the total number of articles in the audience state. This Visibility score is the arithmetic mean of these 146 audience state-level scores. **Visibility_GlobalShare**: The variable represents the percentage [0-100] of media visibility of the analyzed state on the sum of media visibility to all the 147 states in the dataset. It is calculated as the analyzed states’ Visibility score, divided by the sum of all 147 states’ Visibility scores. The scores for all states thus add to 100. ***Use of the data*** --------------------- If you use the data, please cite this site as the source of the data. The Citation section offers several citation styles. This data is being used by the GLOWIN team in a range or work-in-progress article projects. It has been used in the following published or forthcoming work: - Parizek, Michal. ‘Less in the West: The Tangibility of International Organizations and Their Media Visibility around the World’. Forthcoming in *The Review of International Organizations*. A slightly modified version of the dataset has been used in: - Rauh, Christian, and Michal Parizek. ‘Converging on Europe? The European Union in Mediatized Debates during the COVID-19 and Ukraine Shocks’. Forthcoming in *Journal of European Public Policy*, 2024. https://doi.org/10.1080/13501763.2024.2344849. - Parizek, Michal. ‘Worldwide Media Visibility of NATO, the European Union, and the United Nations in Connection to the Russia-Ukraine War’. *Czech Journal of International Relations* 58, no. 1 (2023): 15–44. https://doi.org/10.32422/mv-cjir.60. [1]: https://glowin.cuni.cz/