Predictive analytics have increasingly become an important tool for police agencies in their operations. They have been used as a way of managing decision making that addresses the core functions of law enforcement, including prevention, deterrence, and order maintenance. However, with easy access to data that monitor all aspects of police activity including interactions with the public, there has been an increased pressure on agencies to be more transparent and accountable in their actions. Police have been encouraged to develop protocols that introduce a higher level of procedural justice into their training and join with community members in planning strategies for better relations with the public. In developing this new paradigm, law enforcement has come to rely on crime analysts and criminologists to assist them in managing, analyzing and implementing data to assist in the decision making process.
Difficulties in managing predictive analytics derive from the tension that develops between intuition and evidence based practice. Both feed what is called the risk narrative process. Forming risk narratives is a complicated task. First, we need to consider what constitutes historical fact and what allows certain acts and beliefs to persist. Then, we need to examine how these facts can be transformed into data to be validated, consolidated and aggregated to form the bases for future plans and actions. This leads us to a fundamental question of how this information comes to be filtered through the prisms of intuition, past practice, outcomes assessment, and scientific judgment. The aspects of managing and processing the data are the next step in narrative formation as they help frame problem solving and forecasting. In this context, we need to differentiate automatic responses, based on previous experience, from innovation that made use of data to support new plans and future action. There are, of course, different types of facts and corroboration of these facts is often determined on the basis of experience but also training, culture and ideology. Facts that form data may not display validity if poorly collected or capriciously connected to one another. The formation of these narratives can have a strong influence on how data are used and the reinforcing nature of the storylines that emerge.
This Research Topic aims to solicit research papers that address the varying components of crime analysis and forecasting that influence police decision making, as outlined above. This will draw on many different disciplines (psychology, criminology, public administration, public health, biology, and economics) and will emphasize how evidence-based approaches that incorporate spatial and temporal data can improve how police build risk narratives that improve their daily decision making processes. Authors are encouraged to address some part of the current debates concerning police use of data, crime analytics technology, aspects of crime forecasting and procedural justice, spatial and temporal risk assessment, community/police relations, crime prevention, and training protocols in assisting evidence based decision making.
In addition, as predictive analytics has come under intense scrutiny from civil rights groups and communities that have been subject to policing practices informed by these products, we also welcome articles in this Research Topic that will also address these issues. More specifically, article submissions are encouraged that discuss the potential harms and ethical dilemmas that can come from algorithm-driven police activity. Papers will offer empirically grounded recommendations for ethically sensitive and effective uses of predictive analytics for violence prevention and public safety.
Topic editors Professor Leslie W. Kennedy and Joel M Caplan are partners/co-founders of Simsi Inc. All other topic editors declare no competing interests with regard to the Research Topic subject.
Predictive analytics have increasingly become an important tool for police agencies in their operations. They have been used as a way of managing decision making that addresses the core functions of law enforcement, including prevention, deterrence, and order maintenance. However, with easy access to data that monitor all aspects of police activity including interactions with the public, there has been an increased pressure on agencies to be more transparent and accountable in their actions. Police have been encouraged to develop protocols that introduce a higher level of procedural justice into their training and join with community members in planning strategies for better relations with the public. In developing this new paradigm, law enforcement has come to rely on crime analysts and criminologists to assist them in managing, analyzing and implementing data to assist in the decision making process.
Difficulties in managing predictive analytics derive from the tension that develops between intuition and evidence based practice. Both feed what is called the risk narrative process. Forming risk narratives is a complicated task. First, we need to consider what constitutes historical fact and what allows certain acts and beliefs to persist. Then, we need to examine how these facts can be transformed into data to be validated, consolidated and aggregated to form the bases for future plans and actions. This leads us to a fundamental question of how this information comes to be filtered through the prisms of intuition, past practice, outcomes assessment, and scientific judgment. The aspects of managing and processing the data are the next step in narrative formation as they help frame problem solving and forecasting. In this context, we need to differentiate automatic responses, based on previous experience, from innovation that made use of data to support new plans and future action. There are, of course, different types of facts and corroboration of these facts is often determined on the basis of experience but also training, culture and ideology. Facts that form data may not display validity if poorly collected or capriciously connected to one another. The formation of these narratives can have a strong influence on how data are used and the reinforcing nature of the storylines that emerge.
This Research Topic aims to solicit research papers that address the varying components of crime analysis and forecasting that influence police decision making, as outlined above. This will draw on many different disciplines (psychology, criminology, public administration, public health, biology, and economics) and will emphasize how evidence-based approaches that incorporate spatial and temporal data can improve how police build risk narratives that improve their daily decision making processes. Authors are encouraged to address some part of the current debates concerning police use of data, crime analytics technology, aspects of crime forecasting and procedural justice, spatial and temporal risk assessment, community/police relations, crime prevention, and training protocols in assisting evidence based decision making.
In addition, as predictive analytics has come under intense scrutiny from civil rights groups and communities that have been subject to policing practices informed by these products, we also welcome articles in this Research Topic that will also address these issues. More specifically, article submissions are encouraged that discuss the potential harms and ethical dilemmas that can come from algorithm-driven police activity. Papers will offer empirically grounded recommendations for ethically sensitive and effective uses of predictive analytics for violence prevention and public safety.
Topic editors Professor Leslie W. Kennedy and Joel M Caplan are partners/co-founders of Simsi Inc. All other topic editors declare no competing interests with regard to the Research Topic subject.
![Sample supervised learning ANN [Multilayer Perceptron (MLP)] to predict type of crime from time and place data.](https://www.frontiersin.org/_rtmag/_next/image?url=https%3A%2F%2Fwww.frontiersin.org%2Ffiles%2FArticles%2F587943%2Ffpsyg-12-587943-HTML%2Fimage_m%2Ffpsyg-12-587943-g001.jpg&w=3840&q=75)
