Service-Centre Development, Accessibility, and Regional Development in Northern Medan, Indonesia: A Scalogram, Gravity, Regression, and SWOT Analysis

Abstract

This study evaluates the hierarchy of service centres, potential interdistrict interaction, and the relationships among service-centre development, accessibility, and regional development in Northern Medan, Indonesia. An integrated quantitative-spatial, cross-sectional, perception-based and strategic design was used to connect facility hierarchy, interaction potential, resident assessments, and policy priorities. The study covered Medan Deli, Medan Labuhan, Medan Marelan, and Medan Belawan. It used official 2024 administrative data and questionnaire data analysed in 2026. Primary data were obtained from 120 residents selected through stratified quota sampling, and secondary data were compiled from official statistics and sectoral reports. Descriptive statistics, scalogram analysis, a population-distance gravity model, three simple linear regressions, and SWOT analysis were applied as complementary rather than interchangeable tools. Medan Deli ranked first in the scalogram (score 19), whereas Medan Marelan ranked lowest (score 13) despite having 191,765 residents. The strongest potential interaction was Medan Deli-Medan Marelan (2,013.44 per 1,000,000), followed by Medan Labuhan-Medan Marelan (1,429.53). Accessibility had the lowest mean score (3.38). Service-centre development was strongly associated with accessibility (Beta = .963, R² = .926, p < .001); accessibility was associated with regional development (Beta = .779, R² = .606, p < .001); and service-centre development was associated directly with regional development (Beta = .797, R² = .635, p < .001). SWOT results placed the study area in Quadrant I. Northern Medan requires a differentiated polycentric strategy: consolidate Medan Deli, expand service capacity in Medan Marelan, strengthen Medan Labuhan’s connector role, and provide function-specific coastal-port services in Medan Belawan. Because the regression evidence is cross-sectional and perception-based, the coefficients indicate associations rather than definitive causation.