Cybersecurity Policy Research Driven by Data

Abstract

To support data-driven national cybersecurity policy formulation, we developed a multidimensional analytical framework focusing on global cybercrime distribution, policy effectiveness, and the correlation between demographic characteristics and criminal patterns. First, we collected authoritative data from multiple sources including the International Telecommunication Union (ITU) and conducted preprocessing. Using a K-means clustering model based on Euclidean distance with Python tools, we revealed the global distribution patterns of cybercrime. Second, by integrating time series analysis with Difference-in-Differences (DID) models, we visualized changes in policy-related indicators before and after implementation, quantified policy impacts, and identified key effective policies. Finally, we constructed a multiple linear regression model incorporating interaction effects to visualize correlations between demographic factors and cybercrime, utilizing statsmodels tools for precise predictions. This study demonstrates how the integration of multi-methods and efficient computation provides scientific decision-making support for optimizing cybersecurity policies and global cybersecurity governance.