Development of decision support tool for clustering urban regional risk based on R-ArcGIS Bridge

Abstract

Cities with dense population are susceptible and vulnerable to natural disasters and man-made accidents, making regional risk analysis thus significant for urban public safety. Although risk maps are able to present the broad characteristics of regional risk spatial distribution, quantitative identification of risk clusters in urban areas is still a challenge. In this context, we design and develop a user-friendly customized ArcGIS add-in in this study. In particular, a novel technical routing is proposed for customizing ArcGIS add-in tool based on R-ArcGIS Bridge. The design, architecture and implementation of the tool as well as its core functional modules are introduced. Moreover, R scripts are developed to implement K-means algorithm and Gap statistics validity index for clustering regional risk, respectively. Based on a case study of a typical urban district in China, we introduce the add-in’s functionalities as well as its related decision-making procedures. Results successfully obtained provide evidence that the tool is able to partition regional risk clusters and estimate the optimal number of clusters in urban areas. The simple loosely-coupled architecture also suggests a promising future for embedding some novel geospatial clustering algorithms to extend its capabilities in the next step. This work offers new insights on promoting future urban regional risk management with the use of GIS and clustering algorithms, and it provides a valuable demonstration on extending and enhancing the analysis capacity of GIS by harnessing its power with the statistical analysis capability of R packages through R-ArcGIS Bridge.