Abstract Mobility is a fundamental service in Smart Cities, as commuting between different urban areas underpins daily life. Accurate visualisation of mobility data benefits not only commuters but also traffic managers and city authorities, enabling the identification of congestion hotspots and dominant traffic flows, and supporting timely, data-driven decisions. We propose a novel framework that integrates meshless approximation using Radial Basis Functions with Laplacian-based graph signal recovery to reconstruct and visualise spatio-temporal traffic data across urban road networks. This approach yields smooth, continuous estimates of speed and volume, even in data-sparse regions, and enables the identification of high- and low-traffic zones, as well as dominant intra-city flows. We further estimate vehicular flux using a discrete formulation of Fick’s diffusion law, which reveals patterns of accumulation and dispersal. A road segmentation strategy enhances spatial resolution by dividing roads into smaller segments, enabling finer interpolation and eliminating artefacts in non-road areas. Our findings also reveal areas where traffic volume and speed are negatively correlated. Overall, our framework combines robust data modelling, estimation, and visualisation techniques to support effective urban mobility planning and decision-making.
Márquez et al. (Fri,) studied this question.