Dispersion spatial autocorrelation and mismatches between need and public transport supply in an intermediate city: San Juan de Pasto, Colombia

Abstract

The correspondence between public transport supply and the distribution of the population is usually assessed through a single reading of each variable and with correlations that ignore the neighborhood relationships between zones. This article characterizes the distribution, spatial autocorrelation and associations of nine socioeconomic, territorial and public transport variables in the 184 transport analysis zones (TAZs) of San Juan de Pasto, Colombia, an intermediate city. The 2018 census, the household origin–destination survey, the register of commercial establishments of the Chamber of Commerce and the scheduled supply of the Strategic Public Transport System (SETP) were integrated, and descriptive statistics, Spearman correlations and univariate and bivariate Moran statistics were applied. High dispersion does not imply spatial clustering: commercial establishments are the most dispersed variable (CV = 246.3%) but the least clustered (I = 0.070), whereas income, the least dispersed, is the most clustered (I = 0.455). Scheduled supply is associated with commercial establishments (ρ = 0.58) and income (ρ = 0.38), and is not correlated with households without a vehicle or with people seeking employment. However, the bivariate analysis shows that zones with more households in strata 1 and 2 are surrounded by lower supply (I = −0.235; −0.398 with household rates), a result that is robust to five specifications, and it locates a core of four zones and a northeastern sector where residential need coincides with lower route coverage. The results provide a baseline for assessing accessibility and territorial equity.