Incremental spatiotemporal flow colocation quotient: a new spatiotemporal association analysis method for geographical flows

Analyzing spatiotemporal associations between different types of geographical flows across multiple scales is crucial for understanding the dynamic relationships between them. However, few spatiotemporal association analysis methods exist for geographical flows. Moreover, this analysis also faces the challenge that patterns at larger scales are biased by the cumulative effects from smaller scales if a series of spatial and temporal thresholds at multiple scales is used. Furthermore, spurious results arise if the joint population distribution pattern is not considered. To address these problems, this paper proposes an incremental spatiotemporal flow colocation quotient (ISTFCLQ), which aims to detect spatiotemporal associations between two types of flows across multiple spatial and temporal distance intervals by considering the joint population distribution pattern. The ISTFCLQ designs both global and local indicators for measuring the overall spatiotemporal association patterns of flows and their local association patterns, respectively. Synthetic data tests verified that the ISTFCLQ can effectively reduce cumulative effects and remove biases from the joint population distribution pattern, identifying the exact scale of association patterns. A case study of Xiamen Island taxi and ride-hailing trip data demonstrated the applicability of the ISTFCLQ in analyzing spatiotemporal competition patterns between taxi and ride-hailing services.