Police-Citizen Interactions and Sentiment: A Time Series Analysis of SPD Body-Worn Cameras (BWCs)

• In Event: Police-Community Relations Challenges and Solutions

Abstract

Law enforcement agencies and the public have different views on the key aspects of policing as well as some of the crucial policy issues facing the country. In the mist of significant citizen unrest in recent years due to perceived injustices in the interactions between police and citizens, many government leaders have depended on the use of body-worn cameras (BWCs) as a means of improving citizen relations. Although emerging research on its utility has been generally promising, no evaluation has been conducted on the influence of police-citizen sentiment and its effect on potential escalation and/or de-escalation of a call for service. This study used Natural Language Processing (NLP) to evaluate the relationship between officer sentiment and its influence on the subject via a time series analysis of Seattle Police Department (SPD) body-worn camera footage, second by second, throughout 2021 in over 104k hours of data. This analysis showed a strong correlation between officer sentiment and its prediction of subject sentiment. Additionally, our findings revealed that a citizen’s sentiment at the initial contact with a law enforcement officers (LEOs) played a substantial role in escalation and/or de-escalation of the interaction. Implications for policy and future research are discussed.

Author

• Luis Alberto Gomez-Ventura, Seattle Police Department