Using the Missingness Analysis framework from statistics and the social informatics theory of the political valence of information and sociotechnical systems, as well as publicly available crime complaint data from a US city with a history of using algorithmic policing systems. This paper seeks to investigate fairness issues in the use of historical data in predictive policing systems from a social informatics lens. Moreover, the paper critically analyses the potential consequences, implications, and limitations of historical data within its multidimensional implementation context. The paper also addresses the sustainability of the continued use of historical data in machine learning and algorithmic policing.

Abstract Only