关键词: 聚合型地理流, 时空关联, 双变量时空莫兰指数, 时空权重

Abstract:

Aggregated flows can reflect the spatio-temporal interactions between different spatial areas from the group perspective, and often contain non-spatial attributes that describe quantitative characteristics, such as the migratory intensity of population flows. Analyzing the spatio-temporal association between two non-spatial attributes of aggregated flows can reveal the interaction patterns between different flow attributes, which helps to understand the intrinsic occurrence mechanisms of dynamic interaction phenomena in geographic space. However, there is still a lack of spatio-temporal statistical indicators to evaluate the association degree between two flow attributes, and the understanding of the asymmetry and spatio-temporal heterogeneity of the association remains inadequate. Therefore, this paper proposes a bivariate spatio-temporal association analysis method for aggregated flows. It establishes spatio-temporal weights for aggregated flows to express their spatio-temporal adjacency relationships. Then, it constructs the global and local bivariate flow spatio-temporal Moran's I to assess the asymmetric spatio-temporal association degree between two flow attributes, and identify the local spatio-temporal association patterns and their dynamic variations. The synthetic test results verify that the method can effectively uncover the global and local spatio-temporal association patterns between two flow attributes, and the appropriate scale of spatio-temporal association analysis is also identified through parameter sensitivity tests. The practical application results reveal the spatio-temporal associations between intercity search behavior and travel activities in Shandong, which can provide a theoretical basis for in-depth analysis of urban attractiveness from virtual and real perspectives.

Key words: aggregated flow, spatio-temporal association, bivariate spatio-temporal Moran's I, spatio-temporal weight