Network-constrained Bivariate Clustering Method for Detecting Urban Black Holes and Volcanoes. GeoComputation 2019

Urban black holes and volcanoes are typical traffic anomalies in cities. The discovery of urban black holes and volcanoes has played an important role in urban planning and public safety. It is still challenging to detect arbitrarily shaped urban black holes and volcanoes considering the network constraints with less prior knowledge. In this study, a network-constrained bivariate clustering method is proposed to detect statistically significant urban black holes and volcanoes with irregular shapes. First, an edge-expansion strategy is used to construct the network-constrained neighbourhoods without the time-consuming calculation of the network distance between each pair of objects. Then, a network-constrained spatial scan statistic is constructed to identify candidate sub-areas of urban black holes and volcanoes, which are then combined to form arbitrarily shaped urban black holes and volcanoes based on the multidirectional optimization method. Finally, the statistical significance of each detected urban black hole and volcano is evaluated using Monte Carlo simulation. The simulations demonstrate that proposed method is more effective and stable than the three state-of-the-art methods in detecting urban black holes and volcanoes. The empirical analysis of the Beijing taxicab spatial trajectory data also shows that the proposed method is useful for detecting the spatiotemporal variations of traffic anomalies.