Urban Informatics for Multimodal TOD Diagnostics: GeoAI-Enhanced Node–Place–Design Assessment of Rail and Bus Transit Nodes in Sri Lanka’s Western Province

Transit-Oriented Development (TOD) is a cornerstone of sustainable urban planning. Yet, operationalizing TOD requires urban science methods that integrate heterogeneous spatial data, quantify multidimensional performance, and reveal system-wide patterns. From a complexity-informed perspective, station areas function as complex adaptive systems where accessibility, land use, walkability, and socio-demographic demand co-evolve. In Sri Lanka’s Western Province—where railway stations and major bus terminals structure everyday mobility—TOD decisions are often constrained by transport-centric frameworks and simplified indicators that insufficiently represent multimodal realities and rapid urban change.

This study develops an urban informatics pipeline—integrating big-data, sensing, and computing—to diagnose TOD potential across 92 railway stations and 25 major bus terminals using a GeoAI-enhanced Node–Place–Design framework. The Node–Place Model provides the core analytical theory for assessing the balance between transportation performance (node) and surrounding development intensity/function (place), and is extended with a design dimension to better capture pedestrian accessibility and built-form conditions around transit areas. The assessment integrates 39 indicators operationalized under the 7D TOD principles (Density, Diversity, Design, Destination Accessibility, Distance to Transit, Demand Management, and Demographics). Indicators are normalized, weighted using an entropy-based scheme, and aggregated into a composite TOD index for each node. K-means clustering classifies transit areas into TOD typologies and identifies node–place mismatches (e.g., strong node/weak place). At the same time, an XGBoost model estimates the influence of indicators to highlight priority levers for intervention.

Results reveal substantial spatial variation and persistent imbalances between accessibility and development across the multimodal network, indicating where integrated station-area planning and targeted investments could unlock underutilized transit capacity and support sustainable, equitable urban growth. Theoretically, the study advances urban science by translating a complexity-informed, multimodal Node–Place–Design model into scalable, testable diagnostics; practically, it provides evidence-based typologies and policy levers to target station-area interventions, coordinate land-use–transport integration, and prioritize investments.