Computational Journalism

by

David Caswell

• LAST MODIFIED: 27 June 2022
• DOI: 10.1093/obo/9780199756841-0276

Introduction

Computational journalism is the interpretation of journalistic knowledge in ways that are accessible to algorithmic computation. Journalism interpreted in this way can enable new techniques for gathering, producing, and distributing journalism, or it can deliver new audience experiences of journalism that are not possible when journalistic knowledge is interpreted solely as unstructured natural language. The application of computational approaches to journalism has emerged opportunistically from practice within newsrooms, often using technologies originating from other disciplines and industries, and therefore computational journalism is characterized as a field by a wide variety of techniques, projects, and products without obvious theoretical underpinnings. The practice of computational journalism within newsrooms has primarily been concentrated in three broad categories of automation or computational augmentation; news gathering at scale, the production of news artifacts, and the personalized distribution of news to consumers. Other kinds of computational journalism, such as computationally enabled conversational agents that deliver interactive news experiences, or immersive technologies that deliver more visceral experiences of news, have also received attention from scholars but have proven less sustainable within newsrooms. Computational interventions in the news ecosystem, such as with automated forms of fact-checking or automated monitoring of misinformation and disinformation have also not yet demonstrated utility in routine operation. Academic scholarship of computation in journalism has largely followed the trajectory of the emergence of the field in practice. Early work focused primarily on describing new computational techniques as exotic adjuncts to traditional journalism. As the influence of computation within newsrooms grew, efforts were made to organize the varied expressions of computation in journalism into coherent categorization schemes. Later work recognized the potential for computation in journalism to effect disruptive change and thus became more focused on its impact on audiences and journalists. A small and dispersed technical literature directed at computational forms of journalism also exists, partially within journalism and communication studies, but primarily within computer science, data science, and related fields. Computational journalism’s emergence has occurred within the context of a growing existential threat to journalism and its business models from the effects of ubiquitous networked communication. Computational journalism is therefore often portrayed, in practice and in scholarship, as journalism’s computational response to this computational threat.

General Overview

The relatively recent appearance of computational journalism as a field, combined with its opportunistic emergence from practice, renders it difficult to summarize succinctly. A useful starting point is the relationship between journalism and data, which provides a good basis for appreciating the formation of the field from practice and early scholarship. General reviews of some of the breadth of computational approaches to journalism are valuable, as are more practically focused works describing the actual application of computational techniques within newsrooms. The fundamental nature of journalistic knowledge represented as structured information or data is thoroughly explored by Anderson 2018, culminating from Anderson’s decade-long investigation of the interaction between computation and journalism and grounding computational journalism in pre-digital quantitative interpretations of news. A theoretical treatment of this relationship is available in Anderson 2015, and a technical theory of journalism in computational form is provided by Caswell 2019. Broad descriptions of computational journalism as emerging practice and as a corresponding emerging field of scholarship are available from Thurman 2019 and Caswell and Anderson 2019. An expansive exploration of the application of computation to journalism practice can be found in Diakopoulos 2019, culminating from more than a decade of exploratory work. Thurman, et al. 2021 essentially sums up the first decade of computational journalism as a coherent field, providing a broad range of examples, perspectives, and interpretations. Lindén 2017, although focused on a particular form of computational journalism, provides an accessible and succinct general review of the range of issues surrounding the application of computation within newsrooms. Coddington 2015, in contrast, provides a comprehensive description and analysis of the specific application of end-to-end computational thinking applied to journalism in one nontraditional newsroom. Anderson, et al. 2012 still provides a definitive description of the fundamental crisis facing journalism in the digital era, which is critical context for the emergence of computational journalism.

• Anderson, C. W. 2015. Between the unique and the pattern. Digital Journalism 3.3: 349–363.

  DOI: 10.1080/21670811.2014.976407

  This article explores the idea of abstraction in journalism by contrasting journalism expressed as distinctive reporting with journalism expressed as data-like examples of larger patterns—a concept fundamental to computational interpretations of journalism. Anderson describes these tensions between “story” and “data” and shows how this distinction is a historical one in journalism, predating the introduction of computing machines.

• Anderson, C. W. 2018. Apostles of certainty: Data journalism and the politics of doubt. New York: Oxford Univ. Press.

  DOI: 10.1093/oso/9780190492335.001.0001

  This is a definitive history not only of data journalism, but also of the concept of data journalism, extending from its origins to the experimental frontiers of computational journalism. It traces the expression of journalism as data through several eras, focusing in particular on the interaction between objects of evidence and the social and political environments in which they are interpreted and used.

• Anderson, C. W., Emily Bell, and Clay Shirky. Post-industrial journalism: Adapting to the present. New York: Tow Center for Digital Journalism, 2012.

  Although now somewhat dated, this important work systematically describes the breakdown of the sustaining ecosystem of traditional print journalism as digital networks undermined the distribution barriers of pre-digital news. Although not directly about computational journalism, this report provides a succinct and accessible account of how the chaotic digital news environment in which computational approaches to news operates originated.

• Caswell, David. 2019. Structured journalism and the semantic units of news. Digital Journalism 7.8: 1134–1156.

  DOI: 10.1080/21670811.2019.1651665

  This article provides an analytical interpretation of computational journalism centered on the concept of individual “atomic” elements of news existing on a continuum between natural language artifacts on one end and representation as data elements on the other. This framework is positioned as compatible with similar interpretations emerging in other information-centric domains and is used as the basis of a proposed research agenda for computational journalism.

• Caswell, David, and C. W. Anderson. 2019. Computational journalism. In The international encyclopedia of journalism studies. Edited by T. Vos and F. Hanusch. Hoboken, NJ: Wiley.

  This entry provides a history of the emergence of computational forms of journalism from its earliest origins, as well as a history of the corresponding emergence of academic scholarship of computational journalism. These histories are combined with an examination of the tensions within computational journalism, including between practice and scholarship, and between computational journalism and traditional language-centered forms of journalism.

• Coddington, Mark. 2015. Telling secondhand stories: News aggregation and the production of journalistic knowledge. PhD diss., Univ. of Texas at Austin.

  This PhD dissertation describes in great detail the operations and attitudes of Circa—a remarkable structured journalism news start-up that operated during the first half for the 2010s. Although nominally focused on the aggregation aspects of Circa’s work, the study also provides considerable insight into a novel interpretation of computational news executed in practice by a team of journalists and technologists, including their assumptions and values.

• Diakopoulos, Nicholas. 2019. Automating the news: How algorithms are rewriting the media. Cambridge, MA: Harvard Univ. Press.

  DOI: 10.4159/9780674239302

  This book is a systematic description of all of the major aspects of computational journalism, including news gathering from data, automated production, and automated distribution, grounded in “hybrid” workflows of algorithms and journalists. The book also includes comprehensive chapters on newsbots, on the journalistic reporting of algorithmics, and on future directions in computational news.

• Lindén, Carl-Gustav. 2017. Algorithms for journalism: The future of news work. Journal of Media Innovations 4.1: 60–76.

  DOI: 10.5617/jmi.v4i1.2420

  This essay, based on case studies from three newsrooms, provides an unusually tangible examination of the implementation of computational journalism in practice and directly contrasts common assumptions of the effects of computation on journalism with direct observations. Although an academic study with specific and very useful conclusions, the essay also offers one of the most pragmatic views of how computational journalism is actually interpreted and operated within sophisticated newsrooms.

• Thurman, Neil. 2019. Computational journalism. In The handbook of journalism studies. 2d ed. Edited by Karin Wahl-Jorgensen and Thomas Hanitzsch, 180–195. New York: Routledge.

  This chapter traces the emergence of computational forms of journalism in considerable detail. It concentrates on the emergence of computational journalism as a distinct field, including the emergence of terms and definitions, and provides descriptions of each of the major categories of computation in journalism, including algorithmic accountability. The chapter also reviews different academic interpretations of computational journalism, and integrates the field with the broader landscape of journalism studies.

• Thurman, Neil, Seth C. Lewis, and Jessica Kunert, eds. 2021. Algorithms, automation, and news: New directions in the study of computation and journalism. London: Routledge.

  This collection of papers from a similarly titled conference held in Munich in 2018 offers a broad collection of interpretations and perspectives on computation in journalism, and is one of relatively few collected resources that highlight the emergent nature of scholarship in computational journalism and its attempts to coalesce into a distinct field.