CHOI.J.Y, PARK.M.Y, Kang.J.E. (2023). Assessment of Busan City Central Area System and Service Area Using Machine Learning and Spatial Analysis. ACSP 2023 Annual Conference. (2023. 10. 18. - 2023. 10. 21.)
The balanced spatial structure system increases accessibility to central place that provide various services and improves citizens' convenience and inclusiveness by resolving regional inequality. Recently, Busan City presented the 10CORE, a new urban spatial structure system, in terms of spatial democracy that guarantees an equal quality of life for all in the 2040 Busan City Master Plan. In addition, it is pursuing a ‘15-minute living city’ so that various city functions can be enjoyed within 15 minutes by foot or bicycle from the residential area.
Accordingly, this study aims to identify the existing Busan central place system and evaluate its equity so that Busan citizens can access the equitable central service. Furthermore, by deriving underprivileged areas, it is intended to provide policy implications for future spatial structure planning in Busan.
First, in order to identify the central place, the centrality index was calculated in grid units using four indicators representing the center characteristics. Through Hotspot Analysis, the central place where a high centrality index forms a cluster was identified, and delimiting central area through DBSCAN, machine learning cluster techniques. According to the formula of the center system based on the transportation principle in W.Christaller's central place theory, we can calculate 12 appropriate centers when considering 16 districts in Busan as complementary areas. Accordingly, DBSCAN results representing 12 clusters were selected as the final center. The identified central place are Seomyeon, Dongnae, Sajik, Pusan National University, Yeonsan, Jungang, Busan Station, Daeyeon, Haeundae, Jangsan, Sasang, and Deokcheon.
Interaction indices were used to confirm whether the identified centers had a balanced distribution. Based on the traffic volume between centers, the result of Entropy Index(EI) appeared close to 0 at the level of 0.309, showed an imbalance network between central places. Therefore, we estimated Directional Dominance Index(DII) and it confirmed that central places had a distinct hierarchy. As a result, Seomyeon and Dongnae had a dominance of 2.6 and 1.7 times the overall average, while Deokcheon and Sasang had a remarkably low dominance of 0.3 times. It implies that Busan has an unequal spatial structure system in which hierarchical centers are spatially unevenly distributed.
In order to recognize underprivileged areas according to the unequal central spatial structure, the influence area of central place was derived using Network Analysis. Each influence area adopted the concept of a ‘15-minute living city’ and was designated as a radius of 1 km, equivalent to a 15-minute walk from the central place. As a result, most of the traditional decrepit residential areas in Busan were underprivileged area. Also most of the Eastern and Western Busan, were found to be disadvantaged areas that are not included in the influence area. This suggests that the current central place system does not cover enough residential area.
This study differs from previous studies in that it uses a machine learning technique called DBSCAN in identifying the center. In addition, since the 10 Cores planned by the city have a large gap in hierarchy, it suggests that it is necessary to consider how to narrow the gap between the existing centeral places. It can help to provide equal central area services for citizens. Furthermore, newly added central places such as Gijang, Gangseo, Hadan, and Gadeok can be helpful for balanced development in Busan, but old residential areas in the central region were also needed to be considered in the new central area system in 2040 Busan Master Plan. This study can provide great significance in that it can present the right direction for the equitable spatial structure to the city planning policymakers of Busan.
Keywords: Equity of urban space, Central place, DBSCAN, Interaction Indices, Service area Analysis
사사 : 기후탄력성, 스마트시티