Structured Journalism and the Semantic Units of News

The growing influence of computation on news has intensified the need for an analytical framework that describes the common foundations of different computational approaches to journalism. This article proposes such a framework, founded on the concept of units of journalistic knowledge smaller than the article, expressed partially or completely as structured data and positioned along a continuum of news artifacts. The emerging practice of structured journalism and its use of "atomized" news is described and is presented as an embodiment of the framework. Different approaches to the use of computation within journalism are then positioned as specific instances of that practice. This conception, analogous to "semantic unit" paradigms currently emerging in other information-centric domains, is then used to reinterpret several of journalism's urgent problems. A research agenda for developing computational journalism as an editorial activity within an increasingly data-centric communication environment is proposed, and several implications of the conception are discussed.