The need to enrich the semantic completeness of OpenStreetMap (OSM) data is crucial for its effective use in geographic information systems and urban studies. Addressing this challenge, our research introduces a novel hierarchical feature augmentation approach to developing machine learning classifiers by the features retrieved from various levels of road network connectivity. This method systematically augments the feature space by incorporating measure values of connected road features, thereby integrating extensive contextual information from the network hierarchy. In our evaluation, conducted across diverse urban landscapes in six cities in Italy and Türkiye, we tested two geometry-, six centrality-, and eight semantic-based features to predict road functional classes stored as a highway = * key in OSM. The findings indicate a marginal impact of geometric features and city identifiers on classification performance. Utilizing centrality attributes alongside semantic features in a direct, non-hierarchical manner results in an F1 score of 80%. However, integrating these features in our network-based hierarchical feature augmentation process remarkably increases the F1 score up to 85%. The success of our approach underlines the importance of network-based feature engineering in capturing the complex dependencies of geographic data, considering a more accurate and contextually aware OSM classification framework.

Production models’ flexibilization in capitalist economies continues to transform industrial activities’ spatial organization in a regional continuum. Placed in planned complexes located on cities’ fringes, firms often stand inaccessible from regional circulation routes, which hinder activities’ long-term economic sustainability. Further changes are impending, as forthcoming Smart Manufacturing logistics require efficient linkages between local and regional transportation models. Such issues compel urban planners, economists and policymakers to re-evaluate industrial territories’ imprint on metropolitan dynamics and enact proper strategies towards the industry. In this paper, the role of road-circulation network centralities on industrial complexes’ placement in a regional continuum is analysed, refining the existent methods to assess industry spatial configuration and agglomeration logics. Empirical cases comprise five Brazilian industrial complexes in Porto Alegre’s Metropolitan Region. Hypothesis is that road-circulation network centralities’ hierarchies (closeness and betweenness) have positive correlations to industrial placement patterns at regional and inner-complex scale, informing regional contiguity dynamics amid discontinuous industrial spaces.

Regional configuration can reveal important aspects about city sustainability, as local-regional interactions shape the evolution and inner geography of urban settlements. However, modelling these large-scale structures remains a challenge, due to their sheer size as physical objects. Despite recent improvements in processing power and computing methods, extensive time periods are still required for ordinary microprocessors to model network centralities in road-graphs with high element counts, connectivity and topological depth. Generalization is often the chosen option to mitigate time-constraints of regional network complexity. Nevertheless, this can impact visual representation and model precision, especially when multiscale comparisons are desired. Tests using Normalized Angular Choice (NACH), a Space Syntax mathematical derivative of Betweenness Centrality, found recursive visual similitudes in centrality spatial distribution when modelling distinct scaled map sections of the same large regional network structure. Therefore, a sort of homothetic behavior is identified, since statistical analyses demonstrate that centrality values and distributions remain rather consistent throughout scales, even when considering edge effects. This paper summarizes these results and considers homotheties as an alternative to extensive network generalization. Hence, data maps can be constructed sooner and more accurately as “pieces of a puzzle”, since each individual lesser scale graph possesses a faster processing time.