Kang, YunfanLyu, FangzhengWang, Shaowen2024-08-072024-08-072024-07-17https://hdl.handle.net/10919/120881Network-constrained events, including for example traffic accidents and crime incidents, are widespread in urban environments. Understanding spatial patterns of these events within network spaces is essential for deciphering the underlying dynamics and supporting informed decision-making. The fusion and analysis of networkconstrained point data pose significant computational challenges, particularly with large datasets and sophisticated algorithms. In this context, we introduce NetPointLib, a computationally efficient library designed for processing and analyzing large-scale event data in network spaces. NetPointLib utilizes the capabilities of highperformance computing (HPC) environments including ROGER supercomputer, ACCESS resources, and the CyberGISX platform, providing a scalable and accessible framework for conducting network point data fusion and pattern analysis and supporting computational reproducibility. The library encompasses several algorithmic implementations, including the network local K function and network scan statistics, to enable researchers and practitioners to identify spatial patterns within network-constrained data. This is achieved by harnessing the computational power of HPC resources, facilitating advanced spatial analysis in an efficient and scalable manner.application/pdfenCreative Commons Attribution 4.0 InternationalArticle - Refereed2024-08-01The author(s)https://doi.org/10.1145/3626203.3670615