Proponents of data-driven policing strategies claim that it makes policing organizations more effective, efficient, and accountable and has the potential to address some policing social criticisms (e.g. racial bias, lack of accountability and training). What remains less understood are the challenges when adopting data-driven policing as a response to these criticisms. We present results from a qualitative field study about the adoption of data-driven policing strategies in a Midwestern police department in the United States. We identify three key challenges police face with data-driven adoption efforts: data-driven frictions, precarious and inactionable insights, and police metis concerns. We demonstrate the issues that data-driven initiatives create for policing and the open questions police agents face. These findings contribute an empirical account of how policing agents attend to the strengths and limits of big data's knowledge claims. Lastly, we present data and design implications for policing.
