ZoopZoop: An Integrated Personalized Recommendation and Notification System for Public Policy Information Services

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

Public policy information is often distributed across multiple government institutions and service platforms, making it difficult for citizens to identify policies that match their individual circumstances, interests, and application schedules. Conventional keyword-based policy search requires users to actively identify relevant policy terms and repeatedly monitor policy information, which may result in missed opportunities and inefficient information exploration. To address these limitations, this paper presents ZoopZoop, an integrated personalized public policy information service that combines policy search, personalized recommendation, policy management, deadline notification, and conversational AI assistance within a unified web-based platform. The proposed system collects public policy information through government public data APIs and periodically synchronizes policy lists, detailed information, and eligibility conditions. Personalized recommendations are generated using two complementary sources of information: user behavioral information, including recent search and policy-view histories, and user profile information, including age, income range, residential area, employment status, gender, household size, and marital status. The system further provides notifications for approaching deadlines, newly registered policies, and recommended policies. To support natural-language interaction, an AI chatbot retrieves relevant policy information from the internal policy database and uses the retrieved information as contextual input for response generation. The system was implemented using React, Spring Boot, PostgreSQL, JWT-based authentication, and external public data and AI APIs. The implemented prototype demonstrates how personalized recommendation, proactive notification, and conversational interaction can be integrated into a unified public policy information service. The proposed system provides a practical foundation for improving the accessibility and personalization of digital public policy information services.