Track 5.4: Governance and Management for Real Impact: Embracing Innovation and Managing Emotions

The impact of urban governance quality on residents' quality of life has become increasingly significant. Amid the close relationship between urban sentiment and governance management, studying their interactions is crucial for building harmonious cities. This research investigates how key determinants in urban governance and management influence the generation, variation, and distribution of urban sentiment, aiming to clarify the intrinsic connections between multidimensional governance capacities and urban emotions. The study provides theoretical support and practical guidance for enhancing governance efficiency and optimizing the urban emotional ecosystem. Integrating the “eyes on the street” theory, spatial justice theory, and social capital theory, the research constructs a comprehensive governance evaluation framework. Focusing on Beijing, China, as a case study, the study assesses urban governance capacity across three dimensions: spatial governance, environmental resilience, and social vitality. At the governance level, metrics such as crime spatiotemporal distribution datasets and government open data quantify governance capabilities of specific urban areas. For urban sentiment analysis, the study compiles 555,045 social media posts (Weibo) from 2000 to 2023, utilizing Natural Language Processing (NLP) models to calculate sentiment scores. Ordinary Least Squares (OLS) regression and Geographically Weighted Regression (GWR) models link urban governance to urban sentiment, identifying key influencing factors. Findings reveal that in Beijing, the three governance dimensions—spatial governance, environmental resilience, and social vitality—exert distinct and complex effects on urban sentiment. In the spatial governance dimension, uneven crime distribution directly affects residents' safety perceptions. Areas with high management agency density foster stronger order and security, enhancing positive emotional experiences. Under environmental resilience, park density and accessibility to public facilities emerge as critical factors for happiness. Residents in well-equipped green spaces demonstrate higher life satisfaction and a sense of belonging. In the social vitality dimension, while densely populated areas may cause crowding discomfort, they create more social opportunities and cultural experiences in core regions, strengthening urban identity. Spatially, the central urban area—characterized by rich historical heritage, robust public services, and high governance standards—generates strong pride and belonging among residents, yet also suffers from traffic congestion and high living costs, inducing anxiety. Peripheral regions, though weaker in infrastructure and services, offer lower living costs and relaxed environments, reducing stress levels. This spatial pattern—high satisfaction coexisting with high pressure in cores and low configuration but low stress in peripheries—reflects the deep psychological impacts of urban development imbalances. Theoretically, this study expands the research scope of urban governance and sentiment, deepening understanding of their interplay. Practically, it provides evidence for governments to optimize governance strategies and achieve precise planning. The findings offer insights for addressing challenges such as infrastructure gaps, social inequality, and urban fragmentation, contributing to sustainable urban development.

Around the world, historic towns on the edge of cities are facing an increasingly severe contradiction. On the one hand, there is the huge pressure of development, and on the other hand, there is the urgent need to protect their unique cultural characteristics. In China, the implementation of the "rural revitalization" strategy has undoubtedly made this contradiction more complex and urgent. At this time, social media, like a noisy and real prism, reflects the public's most vivid perception of these spaces. It offers a new lens for understanding these towns through the public's eyes. The constant stream of user-generated content acts as a massive, co-created digital archive full of authentic experiences. Yet, most research has a significant blind spot, analyzing only written comments while ignoring the rich information in photos. This split between text and image may lead to a cognitive bias in the true situation of heritage sites - we hear the story, but miss the scene. The core appeal of this study is to break this single-dimensional information barrier. We try to open up a new interpretation path, by integrating the "narrative" of words and the "evidence" of images, to explore the perceptions and emotions that the public is truly touched by when they step into the historical field. To achieve this goal, this study has constructed a dual-track parallel analysis paradigm. Our data cornerstone is more than 10,000 pictures and text samples from platforms such as Weibo and Ctrip, covering 11 historical towns in Shanghai, and three ancient towns with the most abundant data (Zhujiajiao, Fengjing, and Nanxiang) were selected for in-depth analysis. On the text track, the sentences were disassembled with the help of "jieba", and then a "Erlang Shen-Roberta" model fine-tuned with heritage vocabulary was used to give precise scales to the emotions between the lines, and the modularity algorithm was used to aggregate discrete emotions into clear semantic clusters. On the visual track, we further disassembled thousands of photos into more than 110,000 visual patches, and then used the CLIP model to keenly capture and extract its multimodal features from these fragmented scenes. Thus, we were able to establish a quantifiable mapping relationship between specific landscape elements and the flowing public emotions. The analysis identified four main feature types, with natural spaces and old buildings receiving the highest positive sentiment scores. In contrast, modern facilities scored lower, and visitors' complaints often pointed to poor service. Photo analysis revealed unique visual patterns; for example, water bodies often formed a "water-bridge-pavilion" composition that produced pleasing reflections, while ancient streets presented a "road-wall-top" sequence that provided visitors with a rhythmic walking experience. The integrated graphic and text analysis yielded more nuanced insights. It showed that some frequently photographed building parts evoked weaker emotional responses, which may suggest that they lack deeper cultural significance for visitors. In contrast, rare elements such as fog and bamboo forests, although not frequently seen, were associated with strong positive emotions. These findings suggest that such elements have special importance and deserve greater use in future design and conservation efforts. This research contributes a transferable method that provides actionable insights for planners and managers. It enables a new governance model, shifting from static planning to agile, data-driven management. This approach helps balance heritage preservation with modern needs, fostering a more flexible and inclusive way for these towns to grow.

This paper is initially based on the research done by the author as master thesis research at Politecnico di Milano and Massachusetts Institute of Technology (2020) Innovation beyond innovation districts. Analysis of place-based innovation systems in Boston and Milan. Innovation districts have become major tools for branding and economic growth for cities. However, the definition of the innovation-related concept has changed and evolved through time, which reflects National Innovation Systems, Regional Innovation Systems, and Triple Helix models rooted strongly in the phenomenon. Through the tangible examples of Boston and Milan the primary research explored different governance policy frameworks (Triple, Quadruple helix models, including alternative bottom-up inclusive frameworks) and models in which formal and informal place-based innovation ecosystems emerge. Unintended outcomes and city-wide risks were studied in depth for both cases, where context sensitivity remained the key to success for the innovation ecosystem to succeed. The Current proposal extends this research to Riyadh, Kingdom of Saudi Arabia, where rapid development has started recently through explicit top down approach, governed and managed by entitites like Public Investment Fund, Saudi Data and Artificial Intelligence Authority, Royal Commission, Ministry of Municipal Rural Affairs and Housing, etc, and resulted in multiple mega projects branded as different types of innovation districts such as Diriyah, Digital City, Riyadh Technology Valley, King Abdulaziz City for Science and Technology and Mohamed bin Salman Foundation. With the light of previous in-depth study of Milan and Boston, the research objective is to explore how place-based innovation actually develops in Riyadh, and what we mean when we talk about innovation districts in this context. Governance frameworks, emerging patterns and trends on different levels will be explored closely with the referance to overarching (highly functional and implemented) policy documents such as Saudi 2030 vision, National Industrial Development and Logistics Platform, National strategy for Data & Artificial Intelligence, National Urban Strategy, Research Development and Innovation Program, National Strategy, National Smart Cities Program and different climate initiatives. Research methodology for this study is desk research and comprehensive analysis of scientific literature and the evolution of research debate around the topic. Qualitative and Quantitative tools will be used to analyse and interpret collected data. Desk research is extended to questionnaries distributed to the wider audience of stakeholders and in-person interviews with involved stakeholders, such as local C-level officials, end users, and different social groups. This development that Saudi Arabia witnesses is unusual to how processes are managed in most parts of the world due to the availability of investments, stakeholder relations, and capacity to fully implement the vision documents in practice. On the other hand, social layers and cultural context in Riyadh remain fundamentally different than elsewhere which makes the case of Riyadh most interesting and creates a major condition for the city to grow into its own unique identity. While Riyadh remains to be on a spotlight of worldwide attention for its ambitious initiatives, observing its rapid growth from different angles allows to exploring the direction and potential for innovation in the region. The case of Riyadh reveals a distinct innovation ecosystem shaped by an accelerated, top-down development model rooted in national policy directives rather than organic, bottom-up emergence. Preliminary findings show that while the city's innovation districts benefit from unprecedented financial backing and clear policy alignment, their evolution often lacks participatory governance and contextual adaptability. This could result in challenges such as misalignment with local socio-cultural dynamics, limited grassroots engagement, and performative innovation narratives. However, controversially, the Riyadh model also highlights a new paradigm: one where state-led orchestration, when combined with adaptive governance and inclusive feedback mechanisms, could redefine innovation ecosystems in the Global South.

This applied research project introduces and tests PLANSUSTAIN, an integrated evaluation model designed to assess and enhance sustainability performance in territorial planning. Grounded in rational, communicative, and collaborative planning paradigms, the model bridges conceptual and operational gaps in existing evaluation frameworks by combining multi-criteria sustainability indicators, stakeholder participation, and iterative feedback loops. Through empirical application to mexican urban development instruments, the model revealed how fragmented mandates, vague objectives, and siloed implementation hinder transformative outcomes. Supported by recent funding, the project is now entering a second phase to scale PLANSUSTAIN into an interactive software platform—providing municipalities with a practical, adaptive tool to guide real-time planning decisions and foster cross-sectoral accountability. This presentation shares key insights from the pilot application and outlines the roadmap for tool development and future testing across diverse urban contexts.

Land use diversity plays a critical role in promoting inclusive, resilient, and sustainable urban environments. However, in practice, urban renewal projects often face significant challenges due to competing interests among stakeholders and the high cost of participatory planning processes, such as large-scale surveys, interviews, and multi-round consultations. These constraints hinder the full representation and integration of stakeholder perspectives, particularly in complex redevelopment scenarios. While traditional simulation approaches such as multi-agent-based models (ABM) and cellular automata (CA) have been widely used to analyze spatial and socio-economic dynamics in cities, they fall short in capturing the nuanced cognitive and communicative dimensions of stakeholder interactions. This study introduces an innovative simulation framework that integrates multi-agent systems with large language models (LLMs) to enhance stakeholder negotiation processes in land use planning. The proposed framework is applied to five representative redevelopment scenarios in Bukit Batok, Singapore, an area currently undergoing urban transformation. In this framework, five key stakeholder groups including government agencies, private investors, residents, landowners, and expert opinion leaders are modeled as intelligent agents, each endowed with distinct preferences, constraints, and negotiation strategies. Large language models are employed to simulate realistic dialogue and argumentation patterns based on stakeholder profiles, allowing for dynamic, iterative negotiation processes. To quantify divergence in stakeholder perspectives, we introduce an “opinion entropy” metric, representing the heterogeneity and volatility of expressed preferences throughout the negotiation process. Our results indicate that while governmental and developer agents demonstrated relatively high flexibility in adjusting preferences (entropy index e > 1.4), resident agents exhibited more consistent and rigid demands (entropy index e < 0.6). Despite these initial differences, the iterative negotiation process guided by the LLM-based framework facilitated the convergence of stakeholder interests into a viable, mixed-use land allocation scheme. Compared with the baseline planning scenario, the final simulation output achieved a 48% improvement in functional diversity, reflecting more integrated residential, commercial, and community spaces. These findings suggest that integrating large language models into agent-based simulations can significantly improve the representation of stakeholder cognition and communication in urban planning. The framework offers a scalable and adaptive approach to support participatory decision-making, enabling planners and policymakers to explore a broader range of consensus-based development strategies. By advancing the simulation of negotiation and cognitive conflict resolution, this research contributes to more inclusive, transparent, and adaptive urban transformation processes.

In the era of global co-creation and artificial intelligence, urban governance requires adaptive and forward-looking models to address challenges such as infrastructure, inequality, and social division, and to promote sustainable development. Resource-based cities are confronted with problems such as single industries and great pressure for transformation, and there is an urgent need to optimize urban governance and management. Developing new quality productive forces is an effective measure to promote the transformation and upgrading of resource-based cities. This paper constructs a long-term continuous spatial index system of new quality productivity, including three parts: output, environment and service. Collect data such as patent applications of strategic emerging industries, night lighting data and service facilities in Ma 'anshan City from 2014 to 2023. Use spatio-temporal hotspots and Xgboost-SHAP analysis to explore the evolution characteristics and influence mechanism of the new quality productivity space in Ma 'anshan City. Based on the analysis of spatio-temporal hotspots, it is found that the agglomeration and spillover characteristics of the new quality productive forces in the spatial main urban area of Ma 'anshan City are significant. The space of new quality productive forces has obvious characteristics of spatio-temporal hotspots and can be divided into two types: rapid development and slow development. Through Xgboost-SHAP analysis, it was found that the new quality productivity environment has a significant impact on output, and the influence of new quality productivity services shows an upward trend. Finally, for resource-based cities, it is proposed to lay out a spatial model of "main urban leading and multi-cluster development", strengthen the transformation and development of new quality productivity space in mining areas, and give priority to the development of new quality productivity environment to promote output. Help cities adapt to the demands of the co-creation era, achieve sustainable development, and enhance governance efficiency.