How is Intraday Metro Ridership related to Station Centrality in Athens, Greece?

Abstract

In this study, intraday correlations between station centralities and ridership at stations of the Athens metro system in Greece are explored. An unweighted L-space representation of the physical metro network is developed, and degree, closeness and betweenness are selected as station centrality measures. Hourly smart-card data are used for representing passenger flows. For station classification, principal component analysis and k-means clustering are utilized. The findings suggest that centrality and ridership usually move in opposite directions, morning peak-hour boardings are completely uncorrelated with station centrality, and metro stations can be classified as ‘central destinations’, ‘averagely central origins’, and ‘underutilized peripheral stations’.