Description
The words networks show the issues addressed by the websites of the previous protocol in reporting the Belo Monte controversy. It is evident which are the words shared by many websites categories and which ones belong just to one. The size of the words describes the word recurrency, while the position on the map shows how much the word is shared by the websites. It is possible to identify straight away the clusters of exclusive words.
In the English query network, the three most recurring and shared words are “project”, “Brazil” and “construction”, giving the hint that international websites are oriented in giving a more general overview of the controversy for a non-Brazilian public. In the same query, other frequent words are “river”, “indigenous” and “environmental”, indicating that - aligned with the findings of protocol 2.1 - international websites are stressing visibly the socio-environmental issues. It should be noted that the correspondence of the vocabulary used by Governmental Institutions and Social Media and Publishing Platforms in the English query are due to the fact that different websites were reporting the same content.
On the other hand, the most recurring and shared words in the Brazilian query network are “energia” (energy), “Xingu”, “construção” (construction), “obras” (works) and “hidrelétrico” (hydroelectric.) The mostly addressed issues on Brazilian websites are related to energy and to the construction of the dam itself. Interestingly, it is notable that the words “Xingu” (the river on which the dam is being built) and “Altamira” (the closest city to Belo Monte) are often used by the websites, indicating a very strong connection of these entities with the dam.
Overall, every website category has its own specific language to report the issues linked to the dam, that add to the generally shared vocabulary.
Protocol
Starting from the same list of 25 websites for both queries of protocol 2.1., the protocol aimed to analyse the topics of debate and their link to the sources that reported them. All the text present in every website in the list was inserted in Textalyser, automatically filtering out all the words shorter than three letters. Non-relevant words were manually removed, such as articles. The keywords used for the queries (“Belo”, “Monte”, “usina”, “hidrelétrica”, “dam”) were also removed. The 50 most recurring words were saved in a .tsv file for each analysed website, and the data merged in a single dataset together with the website categories used in the previous protocol. The dataset was imported and visualised with Gephi, and graphically enhanced in Adobe Illustrator. A series of bar charts displaying more detailed data for each word was created as integration for the web experience.
Data
Data Source: Textalyser
Timestamp: 09/11/2018
View Data (5Kb)
Two .tsv files collecting nodes and edges for each network. Words recurrence for each website category is included in this dataset.