["Transactions in GIS, Volume 30, Issue 4, June 2026. ", "\nABSTRACT\nTraditional spatial clustering methods often require input parameters that are difficult to determine, and most use Euclidean or Haversine distances, although many real‐world phenomena are constrained by network space. While several methods have been developed to address this problem, the majority rely on precomputing the pairwise shortest‐path distance matrix, which is computationally expensive. To overcome these limitations, we propose NS‐ACUTE, an efficient topology‐aware clustering method that builds on the ACUTE framework and is specifically designed for geospatial data in network space. The novelty of this study lies in eliminating the need to compute the full shortest‐path distance matrix and in providing an efficient solution for merging clusters without computing distances between them, thereby significantly reducing computational overhead. Furthermore, an adaptive radius mechanism enables the algorithm to enhance clustering quality across different datasets automatically. Evaluation across 10 real‐world datasets from OpenStreetMap, ESRI, and Inside Airbnb demonstrates that NS‐ACUTE achieves an average 84% reduction in execution time compared to state‐of‐the‐art strategies while maintaining or improving clustering quality. These results confirm that NS‐ACUTE offers a robust and efficient alternative for geospatial analysis in network space.\n"]