We present a novel method for analysing socio-spatial segregation in cities by considering constraints imposed by transportation networks. Using a multilayered network approach, we model the interaction probabilities of socio-economic groups with random walks and Lévy flights. This method allows for evaluation of new transport infrastructure’s impact on segregation while quantifying each network’s contribution to interaction opportunities. The proposed random walk segregation index measures the probability of individuals encountering diverse social groups based on their available means of transit via random walks. The index incorporates temporal constraints in urban mobility with a parameter, $$\alpha \in [0,1)$$α ∈ [ 0 , 1 ) , of the probability of the random walk continuing at each time step. By applying this to a toy model and conducting a sensitivity analysis, we explore how the index changes dependent on this temporal constraint. When the parameter equals zero, the measure simplifies to an isolation index. When the parameter approaches one it represents the city’s overall socio-economic distribution by mirroring the steady-state of the random walk process. Using Cuenca, Ecuador as a case study, we illustrate the method’s applicability in transportation planning as a valuable tool for policymakers, addressing spatial distribution of socio-economic groups and the connectivity of existing transport networks, thus promoting equitable interactions throughout the city.

Street networks provide an invaluable source of information about the different temporal and spatial patterns emerging in our cities. These streets are often represented as graphs where intersections are modeled as nodes and streets as edges between them. Previous work has shown that raster representations of the original data can be created through a learning algorithm on low-dimensional representations of the street networks. In contrast, models that capture high-level urban network metrics can be trained through convolutional neural networks. However, the detailed topological data is lost through the rasterization of the street network, and the models cannot recover this information from the image alone, failing to capture complex street network features. This paper proposes a model capable of inferring good representations directly from the street network. Specifically, we use a variational autoencoder with graph convolutional layers and a decoder that generates a probabilistic, fully connected graph to learn latent representations that encode both local network structure and the spatial distribution of nodes. We train the model on thousands of street network segments and use the learned representations to generate synthetic street configurations. Finally, we proposed a possible application to classify the urban morphology of different network segments, investigating their common characteristics in the learned space.

Street networks provide an invaluable source of information about the different temporal and spatial patterns emerging in our cities. These streets are often represented as graphs where intersections are modelled as nodes and streets as links between them. Previous work has shown that raster representations of the original data can be created through a learning algorithm on low-dimensional representations of the street networks. In contrast, models that capture high-level urban network metrics can be trained through convolutional neural networks. However, the detailed topological data is lost through the rasterisation of the street network, and the models cannot recover this information from the image alone, failing to capture complex street network features. This paper proposes a model capable of inferring good representations directly from the street network. Specifically, we use a variational autoencoder with graph convolutional layers and a decoder that outputs a probabilistic fully-connected graph to learn latent representations that encode both local network structure and the spatial distribution of nodes. We train the model on thousands of street network segments and use the learned representations to generate synthetic street configurations. Finally, we proposed a possible application to classify the urban morphology of different segments, investigating their common characteristics in the learnt space.

We present a novel analytical framework to examine socio-spatial segregation across multiple spatial scales, explicitly leveraging information theory and percolation theory. This framework emphasizes the interplay between regional connectivity and population distribution, which are critical for understanding how spatial inequalities arise and persist in urban regions. Employing the Generalised Jensen-Shannon Divergence (GJSD), this method identifies regions characterized by significant segregation and low connectivity, providing actionable insights for targeted urban interventions. Using Ecuador as a case study, we demonstrate how segregation patterns manifest differently at city, regional, and national scales, underscoring the critical role of multiscalarity in understanding urban inequalities and guiding scale-sensitive policies. This approach not only advances the methodology for studying socio-spatial segregation but also contributes to the broader field by highlighting the importance of considering multiscalar perspectives and connectivity in urban systems.