Computational Social Science: A Thematic Review

The explosion of social digital data and the concomitant increases in computational capabilities along the data analytics pipeline (data acquisition, storage and analysis) impact upon the possibilities and choices for conducting social research. This report examines the emerging research field called computational social science (CSS). The aim of this review is to offer insight into the shape of CSS, its questions and methodologies, and how these relate to and interact with different social science disciplines. Two searches and hand sorting identified 41 of the most highly cited publications. The papers were initially categorised into two main groups of papers: substantive-technical contributions and critical-review contributions. The groups were thematically analysed. As a validation and refinement exercise, a further search identified thirty of the most recent CSS papers, which were also categorised and analysed. The review focuses on the first 41 articles as well as several other relevant articles are discussed that were identified through citations, additional ad hoc searches, and personal conversations. The substantive-technical literature and critical-review literature can each be sub-divided into three groups, and findings from these six groups are described. In the discussion, we draw out points related to interdisciplinarity and potential implications of the findings for engagement research communities.