Detecting Multi-Density Urban Hotspots in a Smart City: Approaches, Challenges and Applications

Abstract

Leveraged by a large-scale diffusion of sensing networks and scanning devices in modern cities, huge volumes of geo-referenced urban data are collected every day. Such an amount of information is analyzed to discover data-driven models, which can be exploited to tackle the major issues that cities face, including air pollution, virus diffusion, human mobility, crime forecasting, traffic flows, etc. In particular, the detection of city hotspots is de facto a valuable organization technique for framing detailed knowledge of a metropolitan area, providing high-level summaries for spatial datasets, which are a valuable support for planners, scientists, and policymakers. However, while classic density-based clustering algorithms show to be suitable for discovering hotspots characterized by homogeneous density, their application on multi-density data can produce inaccurate results. In fact, a proper threshold setting is very difficult when clusters in different regions have considerably different densities, or clusters with different density levels are nested. For such a reason, since metropolitan cities are heavily characterized by variable densities, multi-density clustering seems to be more appropriate for discovering city hotspots. Indeed, such algorithms rely on multiple minimum threshold values and are able to detect multiple pattern distributions of different densities, aiming at distinguishing between several density regions, which may or may not be nested and are generally of a non-convex shape. This paper discusses the research issues and challenges for analyzing urban data, aimed at discovering multi-density hotspots in urban areas. In particular, the study compares the four approaches (DBSCAN, OPTICS-xi, HDBSCAN, and CHD) proposed in the literature for clustering urban data and analyzes their performance on both state-of-the-art and real-world datasets. Experimental results show that multi-density clustering algorithms generally achieve better results on urban data than classic density-based algorithms.