Accessibility, traditionally defined as "the degree of proximity to a specific area or facility," has often been quantified through physical distance. This study, however, redefines accessibility from an individualized perspective, incorporating factors such as transportation modes, travel time, destination attractiveness, personal characteristics, and travel purposes. This approach aims to provide a comprehensive metric that transcends conventional distance-based measures, with implications for urban planning, transportation, real estate, and other fields.
Utilizing SK Telecom’s LITMUS data, which records 365 days of individual movement and residence patterns via smartphone base station connections, the study models accessibility as a function of travel impedance (measured by travel time) and destination attractiveness (determined by building types and local commercial characteristics). The analysis focuses on leisure trips in Seoul, categorizing data by gender, age, and transportation mode (public transport versus private car). To ensure accurate estimation, this study employs XGBoost, a machine learning ensemble technique, to analyze over 4.2 million leisure trips, thereby identifying accessibility patterns across various demographic groups.
The results reveal significant differences: for instance, young women in their 20s exhibit higher accessibility to areas such as Gangnam, Mapo, and Yongsan, which are renowned for their diverse leisure facilities and robust public transport connectivity. In contrast, older men in their 70s demonstrate greater accessibility to traditional markets, such as those in Jongno 3-ga, reflecting their distinct activity preferences. These findings underscore the necessity of tailoring accessibility evaluations to account for individual characteristics and travel behaviors.
This individualized framework offers valuable insights for assessing the impacts of policies such as congestion pricing, new urban rail developments, and demographic changes. By integrating these insights into policy design and urban planning, this methodology supports the development of equitable and efficient mobility solutions. Further refinements and expanded datasets may enhance its potential for guiding targeted policy interventions.