Track 2.1: Urban Innovation in Action: Data-Driven Approaches for Adaptability and Resilience

Urban areas face escalating climate-related risks, with flooding emerging as a critical challenge in both humid and arid regions. Rapid population growth, changing land-use patterns, and expanding urban footprints increasingly shape exposure to extreme rainfall and surface runoff. Riyadh, the capital of Saudi Arabia, exemplifies this condition: a fast-growing, semi-arid metropolis where urbanisation dynamics intensify pressure on existing drainage systems and heighten vulnerability to flash floods. Addressing these challenges requires approaches that integrate digital innovation with forward-looking spatial planning to enhance resilience and adaptive capacity. This study aligns with the congress themes of sustainable urban development, resilience, and digital transformation by demonstrating how artificial intelligence can be meaningfully embedded into spatial planning practice. Using a suite of urban and environmental indicators across four flood-prone years, the research generates high-resolution spatial outputs that identify areas of heightened flood susceptibility and, crucially, explain the underlying drivers of risk. The interpretability offered by SHAP enables planners to understand which factors—such as urban density, landform, or rainfall dynamics—most strongly influence exposure in different parts of the city. By providing both predictive insight and transparent reasoning, the framework supports planners and policymakers in prioritising high-risk areas, shaping adaptive zoning strategies, and identifying opportunities for nature-based solutions or infrastructural reinforcement. The temporal dimension further strengthens anticipatory governance by illustrating how risks evolve alongside urban growth. Ultimately, the study demonstrates how AI-enabled analysis can complement spatial planning processes, equipping rapidly urbanising cities—particularly in arid and semi-arid regions—with actionable, evidence-based tools for climate adaptation and resilient urban development.

The integration of artificial intelligence (AI) into urban planning is reshaping decision-making processes, enhancing predictive modeling, and improving the efficiency of early-stage site analysis. This study examines the application of Autodesk Forma, an AI-driven urban planning tool, to assess the spatial and environmental dynamics of King Fahd University of Petroleum & Minerals (KFUPM) Campus, demonstrating the tool’s capabilities in optimizing site-specific analysis. While AI offers real-time data-driven insights, challenges such as algorithmic bias, reliance on built-in datasets, and contextual misinterpretations highlight the need for expert oversight and hybrid AI-human collaboration in urban planning. The research evaluates AI’s role in microclimate, wind, and sun hours analysis. Autodesk Forma was selected due to its user-friendly interface, ability to eliminate manual data uploads, and alignment with the study’s objective of streamlining urban analytics. The results identify key data anomalies, particularly in temperature exceedance projections, prompting the need for external dataset integration to improve predictive reliability. This study contributes to the ongoing discourse on AI in smart city development, demonstrating its potential in early-stage planning while emphasizing its limitations. The findings underscore the importance of integrating expert inputs to refine AI-generated urban models and improve long-term decision-making strategies. The research aligns with Saudi Arabia’s Vision 2030, offering insights into how AI can support sustainable urban growth in technologically evolving cities. The presentation will discuss key findings, challenges, and practical implications for AI-driven urban decision-making, positioning Autodesk Forma as a case study for broader AI applications in smart city planning. By bridging AI analytics with human expertise, this research advocates for collaborative, adaptable urban planning frameworks that enhance both efficiency and resilience.

The emergence of the low altitude economy is fundamentally transforming urban econ omic structures and citizen lifestyles. However, its enabled 24/7 urban mobility networks, AI driven logistics, and nocturnal economic activities remain constrained by infrastructure limitations. These systems were originally engineered for ground based di urnal paradigms, creating a fundamental mismatch with aerial operational requirements that impedes the sector's full potential realization. This research targets high density built up areas. Integrating population spatial distribution, heterogeneous demand patterns, and existing urban morphological conditions, it develops an optimization framework for facility layout while establishing a novel operational paradigm synthesizing physical digital restructuring. This approach resolves the systemic incompatibili ty between legacy infrastructure and low altitude economic requirements, thereby enabling 24/7 dynamic urban systems and AI driven economic digitization. Employing Nanjing as an empirical case, a 25×25 km spatial domain with 60 300m airspace stratification was delineated. A decision support framework coupling multidimensional big data with AI algorithms was developed, integrating: (1) population distribution, (2) diurnal nocturnal economic dynamics (e.g., nocturnal logistics demand, commuter peak flows), an d (3) built environment attributes. Machine learning optimized the tripartite spatial hierarchy (regional zoning/aerial corridors/nodal deployment) balancing physical feasibility, safety constraints, and 24/7 operational efficiency. This generated contextu ally grounded retrofitting strategies for existing structures, transforming them into intelligent low altitude service platforms. This study proposes a conceptual framework integrating artificial intelligence with the built environment through a three dime nsional development paradigm. The framework seeks to reconcile the infrastructural limitations inherent in historic urban districts with the operational requirements of the low altitude economy. It facilitates sustained urban vitality through three core in novations: (1) an optimized hierarchical spatial system enabled by real time demand response matching to socioeconomic fluctuations (e.g., nocturnal logistics surges and emergency service corridors); (2) non destructive adaptive reuse of existing structure s, transforming them into intelligent service hubs; and (3) parametric AI modules that empower cities to dynamically optimize the balance between operational efficiency and social equity across diverse urban communities. By leveraging near ground airspace, this framework transitions urban development paradigms from congested horizontal expansion towards vertically integrated and synergistic three dimensional growth. This activates pervasive digital service provision via low altitude networks. Significantly, it presents an alternative developmental trajectory for mature urban cores seeking enhanced integration with emerging aerial industries, thereby offering a replicable implementation pathway for established urban areas globally to achieve intelligent, 24/7 low altitude economic integration.

The strategic positioning of port cities increasingly shapes global urbanization, trade flows, and digital infrastructure. Gwadar, a flagship development under the China-Pakistan Economic Corridor (CPEC), holds strategic importance within the Belt and Road Initiative (BRI) due to its location at the intersection of Asia, the Middle East, and Africa. While envisioned as a future smart port city and economic hub, Gwadar's growth is constrained by the absence of integrated, predictive planning tools capable of handling its socioeconomic, environmental, and spatial complexities. This study responds to that gap by proposing a hybrid forecasting framework that combines AutoRegressive Integrated Moving Average (ARIMA) modeling with Linear Regression techniques. This approach contributes to the digital transformation of urban planning by leveraging AI-based tools for data-driven, real-time policy insights. Drawing from a collected dataset spanning 2015–2024 sourced from government planning reports, environmental assessments, and international development indicators, this research identifies historical development patterns and projects growth trajectories to 2035. Four critical dimensions were analyzed: (1) Demographic and Economic Trends, (2) Urban Growth and Land Use, (3) Environmental Impact, and (4) Social and Security Indicators. The ARIMA model effectively captures temporal dynamics in population and GDP growth, projecting a 50% population increase and a tripling of GDP by 2035. Linear Regression analysis uncovers strong positive correlations between urban land expansion and CO₂ emissions, as well as between deforestation and settlement sprawl. These findings indicate critical sustainability challenges if environmental safeguards are not prioritized within planning interventions. To position this work within existing literature, the study draws methodological parallels with AI-driven urban forecasting models applied in cities like Shenzhen, Dubai, and Mombasa. Unlike traditional static planning approaches or expert-based predictions previously used for Gwadar, the ARIMA model offers dynamic, time-sensitive forecasts, while regression analysis adds interpretative depth regarding connections between different sectors. For example, studies applying ARIMA in comparable urban contexts report forecasting error margins within ±7%, compared to over ±12% for traditional linear projections. This suggests that adopting such AI-based methods may offer improved accuracy and policy relevance for emerging cities like Gwadar. Social dimensions are also critically examined. The study finds a statistically significant relationship between increased development activity and rising displacement and security incidents, highlighting the risks of uneven economic benefit distribution. These risks are intensified by lack of inclusive planning frameworks. The model forecasts heightened potential for social unrest by 2030 unless countered by community centered urban policy and transparent land governance. By merging AI-driven time-series forecasting with correlation-based spatial and social analytics, this study offers a multi-layered tool to identify urban stress zones, ecological tipping points, and social vulnerability hotspots before problems happen. This integrative method supports real-time urban governance and digital scenario planning, aligning with global smart city discourses. It contributes directly to innovations in urban economic management and exemplifies how AI technologies can be embedded within policy workflows for cities in the Global South undergoing rapid transformation. In conclusion, the study presents a replicable, scalable model for digital urban planning in emerging port cities. Its relevance extends beyond Gwadar, offering insights for planners, researchers, and policymakers working at the intersection of AI, infrastructure-led development, and smart city governance. By addressing the methodological gap and offering a comparative advantage over traditional tools, this research affirms the role of digital innovation in shaping resilient, equitable, and data-responsive urban economies.

Background: Urban innovation systems exhibit significant spatial heterogeneity within metropolitan areas, yet traditional planning approaches treat cities as homogeneous entities. This oversimplification obscures how research facilities, enterprise clusters, urban amenities, and transportation networks differentially contribute to innovation outcomes across urban spaces. As digital transformation accelerates urban economic development, planners increasingly need sophisticated analytical tools to understand these complex spatial relationships and design targeted interventions that leverage specific urban assets for innovation-led growth. Research objective, central question or statement or problem addressed: This study investigates the spatial heterogeneity of urban innovation drivers using interpretable machine learning techniques. Our central question asks: Can machine learning decode distinct spatial patterns of innovation drivers within metropolitan areas and reveal systematic relationships that inform strategic urban planning? We examine how different combinations of urban geographical features contribute to innovation performance and whether cities exhibit characteristic innovation signatures based on their spatial configurations. Research and data collection methods: We analyzed 723 high-innovation 3×3km grids across China's top 10 metropolitan areas using an interpretable machine learning framework. The methodology combines XGBoost algorithms with SHAP (Shapley Additive exPlanations) analysis to quantify feature contributions. Our dataset integrates innovation outputs (academic publications, patents, economic performance) with comprehensive spatial variables including research infrastructure, enterprise density, urban facilities, and transportation networks. Hierarchical clustering of SHAP contribution vectors enabled identification of distinct innovation-driving patterns across metropolitan contexts. Main findings and their significance for theory or practice: (1) Interpretable machine learning reveals four distinct innovation-driving patterns: Research-driven pattern where universities and research institutions serve as primary innovation catalysts (exemplified by Beijing's education districts), enterprise-driven pattern characterized by high-tech company clustering and market-oriented innovation (typified by Shenzhen's technology corridors), urban-life-driven pattern where lifestyle amenities attract and retain innovative talent (found in Guangzhou's amenity-rich quarters), and transportation-hub-driven pattern leveraging connectivity advantages for knowledge diffusion (characteristic of Wuhan's regional centers). (2) SHAP analysis reveals that innovation results from multi-factor synergistic effects, with the combination of research facilities, urban service diversity, and road network density contributing over 50% to innovation variance across all four pattern types, demonstrating the critical importance of these fundamental urban infrastructure elements. (3) Cities demonstrate predictable evolutionary pathways through these patterns—typically progressing from transportation-dependent to research-intensive, then enterprise-driven, and ultimately lifestyle-integrated innovation ecosystems. Theoretically, this study advances urban innovation theory by demonstrating that metropolitan innovation follows identifiable spatial patterns driven by multi-factor synergies rather than uniform processes. The interpretable machine learning approach provides new methodological foundations for understanding complex urban-innovation relationships. Practically, the framework offers planners diagnostic tools for assessing local innovation capacity and designing context-appropriate interventions. Cities can benchmark their innovation profiles against developmental peers, identify strategic priorities based on their dominant patterns, and develop targeted strategies that leverage existing spatial assets. This approach enables evidence-based policy transfer between similar metropolitan contexts and supports precision planning for innovation ecosystem development.

Against the backdrop of globalization, urban innovation is vital for regional competitiveness. This study explores how digital technologies and AI are reshaping innovation ecosystems and contributing to the rise of 24-hour cities, focusing on 41 cities in China’s Yangtze River Delta. Leveraging patent co-invention data (2011–2021) and diverse urban metrics, we decode innovation cooperation networks using an explainable machine learning framework. This framework integrates XGBoost, SHapley Additive exPlanations (SHAP), and Partial Dependence Plots (PDPs) to understand how AI-driven analytics and digital infrastructure – specifically internet user (IU) density and nighttime light data (NPP, a proxy for 24-hour urban activity) – impact collaboration and innovation. We aim to bridge traditional urban economics with AI interpretability, revealing how digital transformation, 24-hour city dynamics, and strategic policies can drive inclusive growth and improve urban lifestyles. Our key findings highlight the interplay between technology and the human experience: (1) XGBoost outperformed conventional models in capturing the evolution of innovation cooperation, revealing a shift from sparse networks to dense ecosystems. The Scientific and Technological dimension, particularly the number of colleges and universities (NCU), is a primary driver. For example, exceeding 11 universities significantly boosts industry-academia collaboration, creating more opportunities for residents to participate in the innovation economy and access new knowledge. (2) The Social and Economic dimension reveals a complex interplay. Internet access (IU) acts as a digital catalyst, synergizing with foreign investment (FDI) and public library resources to boost innovation. Nighttime light intensity (NPP) also shows a positive correlation, indicating that a vibrant 24-hour city fosters innovation, offering more diverse leisure, work, and social activities for residents. However, rising labor costs and traditional industrial clustering can hinder digital innovation, highlighting the need for a transition towards a more digitally-driven economy that benefits workers and residents through upskilling and new job creation. (3) We pioneer the use of NPP as a quantifiable metric for 24-hour city vitality. Our analysis shows that a certain level of nighttime activity (NPP > 11.98 units) correlates with increased infrastructure investment efficiency, leading to improved services and amenities for urban dwellers, such as better transportation, public safety, and access to cultural events. However, excessive urbanization can fragment resources and decrease cooperation, suggesting a need for balanced urban development that prioritizes quality of life, green spaces, and affordable housing. The study suggests stage-specific policy frameworks to maximize the benefits of digital transformation for all residents: Core cities should scale AI-powered university-enterprise partnerships with ESG-aligned funding, fostering a more sustainable and responsible innovation ecosystem that addresses social and environmental challenges. Emerging hubs should prioritize smart infrastructure and public-private partnerships to activate nighttime innovation corridors, creating new opportunities for leisure, work, and social interaction, while also ensuring equitable access to these opportunities. For underdeveloped cities, land value capture and data-sharing platforms can facilitate integration into cross-city AI ecosystems, ensuring that all residents benefit from the digital transformation through improved public services and economic opportunities. By embedding 24-hour city dynamics into an explainable AI model, this research offers a blueprint for leveraging digital inclusivity, policy interactivity, and real-time data to foster resilient, innovation-driven economies that enhance urban lifestyles in the AI era. This work not only addresses the conference’s focus on technology-driven urban transformation but also establishes a globally applicable methodology for decoding complex urban systems through the lens of explainable machine learning model.

Riyadh, one of the fastest-growing global metropolises, faces mounting challenges in population growth, infrastructure management, service equity, and urban sustainability. In response, Riyadh Municipality developed the Riyadh Urban Data Centre (RUDC), a comprehensive urban intelligence ecosystem integrating advanced spatial analysis tools, geospatial platforms, centralized databases, and interactive dashboards. RUDC has enabled digital transformation across the municipality by turning fragmented data into operational and strategic insights, supporting Riyadh’s strategic goals for digital transformation, sustainable development, and Vision 2030. RUDC addresses urban digitalization by offering robust platforms that operationalize data for planning and governance. It serves as a backbone for Riyadh’s shift into a 24-hour smart city through real-time spatial analysis, AI-integrated tools, and a unified ecosystem for urban data. The Centre enhances decision-making in urban growth, infrastructure investment, service distribution, and socio-economic monitoring. With over 190 dashboards, 160 indicators, and spatial integration with municipal tools like the 15-Minute Index, RUDC shows how AI and digital platforms support a resilient, inclusive, responsive, and proactive city. RUDC forms the digital foundation of Riyadh’s shift to a 24-hour city, enabling real-time decision-making, performance monitoring, and scenario simulation through urban intelligence. Riyadh Urban Data Centre applies AI and spatial analytics for service gap detection, sustainable transportation planning, demographic forecasting, urban analysis and trends, and benchmarking. At the core of this transformation are four main pillars: urban studies and reports with over 45 analytical reports delivered across municipal departments, applying global best practices to local needs; 190+ interactive dashboards with 70+ indicators, 165 KPIs, and 350+ data points enabling detailed citywide and neighborhood-level analysis; the Urban Geoportal as a spatial platform for operational tasks such as land permit analysis, service coverage, and neighborhood assessment; and the 15-Minute Index via the Madinati App, a mobile tool for public engagement, measuring proximity to essential services and empowering residents with geo-intelligent feedback. This multi-layered platform through Riyadh Municipality recently earned the Geospatial World Excellence Award 2025 in Smart City & Urban Planning during the World Geospatial Forum 2025 held in Spain. RUDC redefines planning practices across strategic, operational, and public spheres by providing a unified data foundation. Its multi-tiered impact includes strategic planning and urban policy by simulating demand scenarios, benchmarking global smart city metrics, offering advanced demographic projections, and enabling policy coherence across infrastructure, transport, housing, and environmental planning. It supports operational efficiency by allowing municipal departments to identify spatial disparities in services, analyze urban growth patterns, and support annual budgeting. Data is used to assess housing density, environmental exposure, infrastructure distribution, and public realm quality through geo-enabled tools. The Urban Geoportal and dashboards provide land and service data integrated into routine workflows, enabling cross-comparison between districts with filters by indicator type, spatial scale, or timeframe. Citizen-centered tools such as the Madinati app and the 15-Minute Index enhance transparency and participatory planning by showing real-time service access across neighborhoods and allowing residents to explore local data and contribute feedback used in planning refinement. RUDC also strengthens cross-sectoral collaboration by serving as a shared decision-support environment across municipal and national government bodies. It enables interoperability of datasets, reduces redundancy, and accelerates project delivery. As a living ecosystem, the Riyadh Urban Data Centre; A Unified Image for Riyadh City has institutionalized digital transformation in governance. It shows how geospatial intelligence and AI can future-proof city management while enhancing transparency, resilience, and sustainability. This case offers a scalable framework for cities pursuing integrated digital planning, backed by global recognition and municipal uptake.