Evaluating big data policing: The development of a comprehensive quantitative and qualitative evaluation framework

• In Event: WG-PLACE 2: Towards an evidence-based model for big data policing (BIGDATPOL)

Abstract

This presentation unveils a quantitative and qualitative evaluation framework tailored to assess big data policing applications in practice. Recognizing the nuanced nature of qualitative assessments in the context of data-driven policing, the qualitative criteria of the framework emphasize a holistic understanding of the criminological impact, ethics, community dynamics, and, additionally, incorporates a comprehensive cost-benefit analysis. It encompasses key dimensions such as the interpretability of data-driven insights, ethical considerations in algorithmic decision-making, and the social and criminological implications of police interactions influenced by big data. It places particular emphasis on the perspectives and experiences of diverse stakeholders, including community members, to capture nuanced insights. In addition to these qualitative perspectives, the evaluation framework pays attention to quantitative aspects in evaluating big data policing applications. These quantitative evaluations cover both statistical (e.g., performance of machine learning models) and substantive (e.g., crime reduction effects) aspects. By offering a qualitative and quantitative lens on big data policing, this presentation aims to provide a comprehensive tool for policymakers, law enforcement agencies, and researchers to assess the multifaceted aspects of data-driven strategies. The presentation concludes with reflections on the potential evolution and refinement of the evaluation framework to meet the evolving needs and challenges of contemporary policing.

Authors

• Marlies Sas, Ghent University

• Thom Snaphaan, Avans University of Applied Sciences & Ghent University

• Wim Hardyns, Ghent University