Visualization, technologies, or the public? A text mining analysis of the articulation of data-driven journalism in the Twittersphere

Data-driven journalism has triggered debates that whether these innovative approaches, such as using data analytical and computational methods, better serve the public. Applying the concept of articulation, wherein an array of terms are juxtaposed and expressed together, this paper examined how the term “data-driven journalism” is discursively constructed by ordinary people on social media. Using the Twitter application programming interface (API), this paper harvested all available public tweets (n = 2,597) containing hashtags or keywords related to data-driven journalism within two weeks in late October 2016. Text-mining indicated these tweets focused intensively on data visualization and data analytical techniques. Further analysis on the hashtag co-occurrence network, i.e., a network established via creating edges between two or more hashtags appearing together within the same tweet, revealed that journalism and visualization-related hashtags were located at important positions in the network; in contrast, public-related terms, such as “#opendata” or “#opengovernment,” were positioned peripherally.