| dc.description.abstract | Measurement of spatial accessibility is useful for the assessment of the ease of access to an infrastructure of interest from the interactions of supply, demand, and mobility. Spatial accessibility measures are essential in the investigation of the geospatial problem, as they explain the spatial disparity of access between inter- and intra-regions and suggest that locations of poor accessibility should be supplemented with additional resources. However, the conventional approach used in spatial accessibility studies underexamined the temporal dynamics embedded in the three input variables, overestimating accessibility and decreasing the accuracy of measurements. To address these concerns, the implementation of temporal dynamics in spatial accessibility is thoroughly investigated in three aspects as follows: 1) examining the stochastic distribution of spatial accessibility from the uncertainty underlying the temporal changes, 2) enhancing the accuracy of accessibility measurements by leveraging the temporal changes of supply, demand, and mobility, and 3) promoting the reproducibility of temporal dynamics by prioritizing the use of time-dependent variables in accessibility measurements.
The analyses performed in this dissertation provided three major findings. First, it revealed that the uncertainty embedded in the temporal dynamics dramatically changes the degree of accessibility and intensifies the inequality of access in the worst-case scenario. This indicates that the temporal changes should be considered in assessing accessibility, given their notable impact. Second, incorporating time-dependent variables into measurements would uncover the 24-hour variation of spatial accessibility. The enhanced temporal granularity of the measurement provides an improved understanding of the time-dependent spatial disparity of access and supports a better allocation of supplementary infrastructures at a specific space and time. Third, the implementation of time-dependent mobility, rather than time-dependent supply and demand, should be prioritized to maximize the temporal variations of the measurements, and the partial use of time-dependent variables may fail to predict accessibility during the daytime. These findings shed light on the benefits of utilizing time-dependent variables and suggest possible strategies to promote policy implications. Therefore, the dissertation would provide insights on how to take advantage of the current data-rich environment to understand the temporal dynamics of geospatial problems for the discipline of Geographic Information Science and Systems. | |
