Abstract
Community is the basic unit of public safety management. A good community safety service is the key to improving community safety levels. In the era of data explosion, big data-based decision-making has been widely adopted to facilitate community safety. This paper proposes an integrated "driving model" for big data governance specifically tailored for urban community safety. By combining scenario-driven, data-driven, and model-driven approaches, the framework moves beyond simple data sharing to solve complex, multi-level governance problems like privacy protection and organizational synergy.
Introduction
As cities evolve into "smart" ecosystems, they generate massive volumes of data essential for monitoring environmental risks and managing public services. However, this data is frequently siloed, inconsistent, or restricted by privacy concerns. Current big data governance research often remains fragmented, focusing either on technical data quality or high-level regulation without integration. For instance, a privacy issue is rarely just a technical encryption problem; it involves organizational regulation and metadata structure. There is a critical need for a systematic model capable of handling these integrative challenges in high-stakes urban safety environments.
The Integrated Driving Model
The core contribution of this research is a three-level Driving Model of Big Data Governance, designed to bridge the gap between abstract policy and technical implementation.
- Scenario-Driven Level (SDL): Focuses on "solution generating." It employs techniques like "Mess Maps"—a question-tracing methodology—to visualize and structure complex social problems that lack direct, linear answers.
- Data-Driven Level (DDL): Provides "objective evidence" through mathematical similarity calculations (e.g., Euclidean and Manhattan distances). This layer compares current safety scenarios against historical data to identify patterns.
- Model-Driven Level (MDL): Offers "process support" via standardized procedures and algorithms. This includes Game Theory for stakeholder coordination and Delphi models for evaluating the effectiveness of implemented solutions.
Case-Based Reasoning and Scenario Matching
A standout feature of this model is the application of Case-Based Reasoning (CBR). Rather than developing new solutions for every safety issue, the system calculates the similarity between the current problem and a database of historical cases to recommend proven governance strategies.
By treating a "governance scenario" as a vector of features—such as data precision, authorization status, and privacy risk—the model uses geometric distance formulas to find matches. If a current safety crisis shares a high similarity (e.g., 85%) with a past event, the previous solution serves as a high-probability starting point for intervention.
Case Study: Urban Position Data
The model was validated using a use case from "S city" in China, focusing on position data extracted from social media. The authors categorized the scenario into 18 detailed feature types to test the CBR technique's ability to match real-world problems to historical solutions.
The system utilized feature weights—assigning higher importance to critical factors like "Authorization Status"—to identify the most relevant historical matches. For the "S city" administrators, the model generated four concrete operational steps: updating the position corpus, improving classification thresholds, sampling, and testing samples. This demonstrated the model's ability to translate abstract big data concepts into actionable safety strategies.
Conclusion
The proposed driving model acknowledges that technical fixes are insufficient without organizational synergy. By integrating game theory into the model-driven level, the framework encourages cooperation among departments that might otherwise be protective of their data. Future work will focus on increasing the model's universality by connecting it to broader reference models and conducting more extensive empirical testing across diverse urban environments and safety emergencies.
Summary updated