Natural language processing in government applications: a literature review and a case analysis

Purpose: This study aims to conduct a comprehensive literature review on the use of natural language processing (NLP) in government applications and analyzes the findings from various perspectives. Drawing from this academic review, we propose a classification framework for understanding the role of NLP in government applications. To provide practical insights, we present a case study focusing on the development of a social media text analytics system in a government department in Hong Kong. This case study documents the design and implementation of the system, showcasing its ability to provide timely supporting information to decision-makers. Design/methodology/approach: In this research, a two-stage research methodology is employed. The first stage involves a thorough literature review of articles focusing on “NLP applications in government.” The second stage comprises a case study of a government department currently utilizing NLP technologies to enhance government services. This case study not only validates the conceptual framework and societal value grid but also illustrates the design, development and implementation of an NLP application system within a government department in Hong Kong, emphasizing the actual value of the system. Findings: The findings of the literature review contribute to the development of a conceptual framework and a societal value grid. The framework outlines major domains where NLP is applied and the key techniques supporting these applications, while the NLP societal value grid provides an overview of NLP’s societal benefits in the government sector. Research limitations/implications: This study contributes by conducting a comprehensive literature review on NLP technologies applied to government applications and formulating a conceptual framework for the classification of NLP technologies to government applications. Additionally, the contributions of this research include the design artifacts – a real-time social media analytics prototype system (comprising both system architecture and implemented prototype). Practical implications: The design and implementation of the system, showcasing its ability to provide timely supporting information to decision-makers. We introduce a social media analytics value grid designed for managers and officers in the government sector, aiming to enhance their understanding of the social media text analytics system and its role in capturing social value. Social implications: The development of a social value grid, incorporating an impact and/or value framework for real-time assessment of the social analytics system’s value. This grid enables benchmarking against other government departments. Governments can utilize NLP and text analysis technologies to mine public sentiment and expert content, gaining insights into citizen preferences and information relevant to policy propositions or implementations. Originality/value: The development of a social value grid, incorporating an impact and/or value framework for real-time assessment of the social analytics system’s value, adds to the contributions. This grid enables benchmarking against other government departments. We believe this study establishes a theoretically grounded foundation for future research to delve into the value of social media sentiment analysis for government applications.