Data-Driven Urban Planning

Helsinki just marked a full year without a single traffic fatality. https://lnkd.in/ec--U4Ar I once cycled through Finland (interspersed with train travel to make it round some of their 187,000 lakes!) From this I became slightly obsessed with travel pattens and transport options across the country. So, coming across this latest headline was great to see, but it wasn’t a surprise as such, it’s been years in the making. And here’s why:   It’s not just about safer crossings or speed limits in isolation. Helsinki’s transport authority uses tools like mobile sensing (e.g. TravelSense) more recently real-time bike-share data, and detailed travel-time matrices to understand the spatial and temporal rhythms of the city. 📍 They know which streets see early-morning cycle spikes. 🚶♀️ They can track when pedestrian volumes rise near rail hubs. 🚌 They model access to schools, services, and jobs by time of day and by mode Data‑driven infrastructure and policy in Helsinki isn’t blind, it's tuned to when and where people move, across different user types and modes. This long term data gathering and understanding is a key driver behind the city’s safety and modal‐shift successes. This insight supports practical decisions—30 km/h limits around schools and residential zones, widened footways and separated bike lanes in high‑footfall areas, and targeted improvements in under-served parts of the city. All this feeds into how people behave and how we use this to shift behaviour. As the article points out, the credit for zero deaths belongs not just to infrastructure, but to everyone using the road. When public space invites care, predictability and mutual respect, behaviour changes too. Better safety isn’t just about enforcement or campaigns. It’s about designing systems that anticipate movement patterns, support good decisions, and allow human error without tragic consequences. 👉 In other words: this is a Safe System approach in practice. Some supporting research for those interested: https://lnkd.in/edFtH2aC https://lnkd.in/eHtwV6wM #SafeSystem #VisionZero #UrbanMobility #RoadSafety #BehaviourChange #CityPlanning #ActiveTravel

📘 Artificial Intelligence in Urban Planning and Design: Technologies, Implementation, and Impacts As cities become more data-driven and complex, AI is no longer a futuristic concept in urban planning — it is an active design force. 🌆🤖 This comprehensive resource explores how Artificial Intelligence is transforming smart city planning and urban design. 🔍 Why This Matters It goes beyond surface-level discussion and provides: → 🧠 A clear foundation of AI theory in the context of urban systems → 🏙️ Real-world applications of AI in city planning and design → 📊 AI-driven research and information systems → 🎨 Generative design frameworks powered by AI Rather than presenting AI as a single tool, it positions AI as a structural shift in how cities are analyzed, modeled, and designed. 🚀 A New Design Paradigm One of the most compelling themes is the rise of AI-generated planning solutions — often created without predefined rules. This introduces powerful opportunities: ✔️ Adaptive urban modeling ✔️ Data-informed infrastructure planning ✔️ Dynamic simulation of growth scenarios But it also raises critical questions: • Who defines the objectives? • How do we ensure transparency? • What happens to traditional planning expertise? 🧩 Theory Meets Practice It bridges: 🔹 Theoretical foundations of AI 🔹 Practical implementation in urban systems 🔹 Critical evaluation of tools and methodologies 🔹 Future directions for responsible AI integration AI is not treated as a silver bullet. Instead, both potential and limitations are examined with balance. 🌍 The Bigger Picture Urban environments are living systems — socially, economically, and environmentally interconnected. AI introduces the possibility of: • More resilient city planning • Optimized resource allocation • Smarter infrastructure design • Human-centered urban innovation Meaningful progress requires thoughtful governance and intentional design. AI in urban planning isn’t just about smarter cities. It’s about designing cities that remain human at scale. Follow and Connect: Woongsik Dr. Su, MBA #ArtificialIntelligence #UrbanPlanning #SmartCities #GenerativeDesign #DigitalTransformation #UrbanInnovation #CityPlanning #AIInDesign

𝐖𝐡𝐲 𝐭𝐡𝐞 ‘𝐓𝐡𝐫𝐞𝐞 𝐀𝐥𝐭𝐞𝐫𝐧𝐚𝐭𝐢𝐯𝐞𝐬’ 𝐌𝐨𝐝𝐞𝐥 𝐢𝐬 𝐊𝐢𝐥𝐥𝐢𝐧𝐠 𝐂𝐢𝐭𝐲 𝐏𝐥𝐚𝐧𝐧𝐢𝐧𝐠 – 𝐀𝐧𝐝 𝐖𝐡𝐚𝐭 𝐖𝐞 𝐒𝐡𝐨𝐮𝐥𝐝 𝐃𝐨 𝐈𝐧𝐬𝐭𝐞𝐚𝐝 For decades, urban planning has followed the three-alternatives model, often leading to a hybrid fourth option—sometimes strategic, but often a reactionary mix of ideas. When done right, alternatives provide flexibility, but when built without data, scenario testing, or probability modeling, they can kill a city’s potential before it even takes shape. 𝗪𝗵𝗮𝘁 𝗚𝗼𝗲𝘀 𝗪𝗿𝗼𝗻𝗴? - Alternatives without scenario-driven foundations lead to fragmented, uncoordinated urban growth. - Decisions based on hybridizing weak ideas instead of selecting the best-tested option. - Lack of probability-based forecasting, making urban expansion a guessing game. 𝟭𝟮 𝗦𝘁𝗲𝗽𝘀 𝘁𝗼 𝗖𝗿𝗲𝗮𝘁𝗲 𝗦𝗺𝗮𝗿𝘁𝗲𝗿, 𝗦𝗰𝗲𝗻𝗮𝗿𝗶𝗼-𝗗𝗿𝗶𝘃𝗲𝗻 𝗔𝗹𝘁𝗲𝗿𝗻𝗮𝘁𝗶𝘃𝗲𝘀 To prevent urban failure, alternatives must be built on data, probability models, and scenario forecasting. Here’s how to do it right: 𝟭) Define Key Drivers Using Probabilistic Analysis – Identify economic, demographic, and climate trends using Monte Carlo simulations and historical data. 𝟮) Set Scenario Time Horizons & Probability Weights – Assign likelihood scores to different futures (Compact City = 60%, Sprawl = 30%, Decentralized Nodes = 10%). 𝟯) Use Bayesian Forecasting for Data-Driven Projections – Refine infrastructure demand and land use forecasts based on real estate and economic trends. 4) Develop Multiple Scenarios with Risk Probability Scores – Avoid single-outcome planning by testing multiple futures under different policy and economic stress tests. 5) Translate Scenarios into Spatial Alternatives – Ensure each alternative directly reflects a tested scenario, not just an arbitrary layout. 𝟲) Test Alternatives Against Economic & Environmental KPIs – Use real estate absorption models, climate risk scores, and probability-adjusted cost-benefit analysis. 𝟳) Factor in Policy & Regulatory Risks – Model zoning law changes, governance shifts, and regulatory enforcement trends to prevent future conflicts. 𝟴) Incorporate Economic Feasibility & ROI Projections – Use discounted cash flow (DCF) modeling to assess long-term financial sustainability 𝟵) Avoid Arbitrary Hybridization—Use Data to Justify Merging Alternatives – Only combine alternatives if probability models show compatibility, not as a political compromise. 𝟭𝟬) Engage Stakeholders & Test Probabilities with Digital Simulations. 𝟭𝟭) Plan Phased Implementation Based on Infrastructure Readiness – Align urban expansion with stochastic forecasting of infrastructure demand. 𝟭𝟮) 𝗦𝘁𝗿𝗲𝘀𝘀-𝗧𝗲𝘀𝘁 𝗦𝗰𝗲𝗻𝗮𝗿𝗶𝗼𝘀 for Black Swan Events – Model low-probability, high-impact disruptions 𝙏𝙝𝙚 𝘽𝙤𝙩𝙩𝙤𝙢 𝙇𝙞𝙣𝙚: 𝙋𝙡𝙖𝙣𝙣𝙞𝙣𝙜 𝙒𝙞𝙩𝙝𝙤𝙪𝙩 𝙎𝙘𝙚𝙣𝙖𝙧𝙞𝙤𝙨 𝙇𝙚𝙖𝙙𝙨 𝙩𝙤 𝙐𝙣𝙘𝙚𝙧𝙩𝙖𝙞𝙣𝙩𝙮 #urban_planning #Urban_design #cityplanning

What if we could redesign planning regions based on the problem we are trying to solve, not on boundaries drawn decades ago? Excited to share our latest publication in Nature Scientific Reports: "Demand-Oriented Regionalization with Local Data for Climate Adaptation Planning" by Mobina Noorani, Shangde Gao, Changjie Chen, and Karla Saldaña Ochoa. link: https://lnkd.in/e2pNdt-f Cities often rely on census tracts, neighborhoods, or administrative boundaries to make decisions about climate adaptation, infrastructure investments, and public health interventions. However, these boundaries rarely align with the actual spatial patterns of risk. In this work, we developed RepSC-SOM (Representative-initialized Spatially Constrained Self-Organizing Map). This AI-driven regionalization framework creates planning regions based on local environmental and socioeconomic conditions rather than predefined administrative units. Using Jacksonville, Florida, as a case study, we integrated: Flood inundation risk Septic system density Property values Impervious surfaces Built environment characteristics Population density to generate data-driven regions tailored to climate adaptation planning. The results were compelling. Our regionalization framework identified flood-related water contamination hotspots more effectively than traditional planning units, such as census tracts, neighborhoods, and traffic analysis zones. Regions generated by RepSC-SOM showed substantially greater differentiation in E. coli contamination levels, enabling more targeted adaptation and resource allocation strategies. Beyond flood resilience, this framework can be adapted to address challenges such as: Extreme heat Air quality Public health risks Transportation planning Environmental management This research advances our vision of human-centered AI for urban resilience, where machine learning helps planners identify meaningful spatial patterns while maintaining transparency and interpretability. A huge congratulations to the team for this achievement! #ArtificialIntelligence #ClimateAdaptation #UrbanPlanning #DigitalTwins #MachineLearning #Resilience #SmartCities #UrbanAI #ClimateResilience #Jacksonville #PlanningSupportSystems #HumanCenteredAI

𝗧𝗮𝗸𝗲 𝘁𝗵𝗲 (𝗔𝗿𝗰)𝗚𝗜𝗦 𝗢𝘂𝘁 𝗼𝗳 𝗦𝗺𝗮𝗿𝘁 𝗖𝗶𝘁𝗶𝗲𝘀 𝘄𝗶𝘁𝗵 𝗙𝗦𝗤 𝗦𝗽𝗮𝘁𝗶𝗮𝗹 𝗔𝗴𝗲𝗻𝘁 Cities collect more spatial data than ever, yet most municipalities can't extract actionable insights from it. Municipal teams either commission expensive consulting firms and wait weeks, or task GIS specialists with wrangling data at different precisions and incompatible formats. The problem isn't lack of data—it's the technical barrier between questions and answers. 𝗙𝗦𝗤 𝗦𝗽𝗮𝘁𝗶𝗮𝗹 𝗔𝗴𝗲𝗻𝘁: 𝗔𝘂𝘁𝗼𝗻𝗼𝗺𝗼𝘂𝘀 𝗦𝗽𝗮𝘁𝗶𝗮𝗹 𝗥𝗲𝗮𝘀𝗼𝗻𝗶𝗻𝗴 𝗳𝗼𝗿 𝗨𝗿𝗯𝗮𝗻 𝗜𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 One prompt: "𝘞𝘩𝘢𝘵 𝘪𝘯𝘴𝘪𝘨𝘩𝘵𝘴 𝘤𝘢𝘯 𝘺𝘰𝘶 𝘨𝘦𝘯𝘦𝘳𝘢𝘵𝘦 𝘧𝘰𝘳 𝘶𝘳𝘣𝘢𝘯 𝘱𝘭𝘢𝘯𝘯𝘪𝘯𝘨 𝘪𝘯 𝘊𝘩𝘢𝘯𝘥𝘭𝘦𝘳, 𝘈𝘳𝘪𝘻𝘰𝘯𝘢?" The agent autonomously explored the H3 Hub and selected eight relevant datasets based on semantic understanding of urban planning domains—WorldPop population density, US Census housing and income metrics, Overture building infrastructure, FEMA flood zones, OpenCelliD cell towers, power transmission lines, and medical services. It identified 251 hexagonal spatial cells covering the entire city, executed joins across datasets that would otherwise be incompatible, built composite scoring algorithms, and explained its reasoning at every analytical decision point. 𝗧𝗵𝗲 𝗥𝗲𝘀𝘂𝗹𝘁𝘀 The agent's urban planning insights for Chandler outline a holistic transformation centered on four interconnected themes (see video for details): 🏙️ 𝗦𝗽𝗮𝘁𝗶𝗮𝗹 𝗘𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝗰𝘆: Shifting from sprawling residential layouts to dense, mixed-use "15-minute city" designs. 📶 𝗦𝘆𝘀𝘁𝗲𝗺 𝗖𝗼𝗻𝗻𝗲𝗰𝘁𝗶𝘃𝗶𝘁𝘆: Bridging severe physical and digital gaps via multi-modal transit and expanded cell networks. 🤝 𝗛𝘂𝗺𝗮𝗻 𝗘𝗾𝘂𝗶𝘁𝘆: Tackling housing affordability and protecting vulnerable populations from displacement. ⚡ 𝗦𝘆𝘀𝘁𝗲𝗺 𝗥𝗲𝘀𝗶𝗹𝗶𝗲𝗻𝗰𝘆: Future-proofing infrastructure with crucial upgrades to power grids and emergency services. 𝗧𝗵𝗲 𝗙𝘂𝘁𝘂𝗿𝗲 𝗼𝗳 𝗚𝗲𝗼𝘀𝗽𝗮𝘁𝗶𝗮𝗹 𝗜𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 𝗳𝗼𝗿 𝗦𝗺𝗮𝗿𝘁 𝗖𝗶𝘁𝗶𝗲𝘀 FSQ Spatial Agent couples reasoning LLMs with geospatial data normalized to the H3 hexagonal grid, making every dataset queryable at the same resolution and eliminating weeks of preprocessing. The agent constructs composite scores with transparent weighting, explaining its reasoning at every decision point. Because everything is H3-standardized, planners can bring in their own municipal data—building permits, 311 service requests, traffic counts—and the agent seamlessly integrates it without code. Traditional workflows demand expensive licenses and weeks of work. FSQ Spatial Agent delivers comprehensive analysis in minutes. 𝗔𝘃𝗮𝗶𝗹𝗮𝗯𝗹𝗲 𝗻𝗼𝘄 𝗶𝗻 𝗙𝗦𝗤 𝗦𝗽𝗮𝘁𝗶𝗮𝗹 𝗗𝗲𝘀𝗸𝘁𝗼𝗽 (𝘋𝘰𝘸𝘯𝘭𝘰𝘢𝘥 𝘭𝘪𝘯𝘬 𝘪𝘯 𝘤𝘰𝘮𝘮𝘦𝘯𝘵𝘴)

Real estate and cities have been planned the same way for decades — static models, fragmented data, slow iteration cycles. That approach is reaching its limits. Generative Urban AI changes the starting point. At SynPlanet, we launched a generative urban system that works with the city as a single environment — across thousands of data layers: terrain, underground networks, infrastructure, mobility, zoning, buildings and market context. Instead of designing one option, teams can explore many viable scenarios instantly. This affects both city operations and real estate decisions. Urban teams can test infrastructure and policy scenarios before committing. Developers can compare multiple project configurations before design begins. Public agencies can understand consequences upfront — not after implementation. The key shift is the combination of generative simulation with a Large Geospatial Model. Scenarios are not only visualized — they can be evaluated. Which option improves mobility? Which increases land value? Where do infrastructure constraints appear? What risks emerge over time? Which scenario performs better operationally? After selecting a scenario, planning documentation can be generated in formats aligned with local city requirements. Concept → evaluation → approval preparation becomes a continuous workflow. This is not a concept stage. The product is live. What is emerging is a new layer in the built environment stack — Generative Urban Intelligence — where cities, developers and infrastructure operators move from designing projects to computing scenarios. #GenerativeAI #UrbanAI #GeospatialAI #RealEstateAI #PropTech #SmartCities #CityPlanning #Infrastructure #BuiltEnvironment #DigitalTwins #SynPlanet

A recent feature in Smart Cities Dive illustrates how the City of New Orleans shifted away from disconnected tracking tools to embrace data-driven #infrastructuremanagement. By integrating real-time telemetry and video AI across 41 separate departments, they've established a highly practical blueprint for municipal modernization. Three operational realities stand out for state and local #governmentleaders: 1. Ditching the Paper Trail for Public #Accountability: When managing thousands of public assets, moving to a centralized, real-time tracking framework gives leadership instantly verifiable visibility into asset location and utilization. 2. Proactive Readiness Over Reactive Repairs: #Telemetry platforms allow fleet managers to monitor engine diagnostics and pending maintenance checks continuously, ensuring vehicles are road-ready before crews deploy. 3. Measurable Safety Improvements via #AI Insights: By utilizing AI-enabled dash cams and behavioral feedback loops, New Orleans successfully reduced driver speeding by 37% and mobile phone distractions by 46% over a 12-month period. True #smartcity transformation isn’t about chasing abstract tech trends or buying hardware for its own sake. It’s about connecting physical operations to a unified data layer to save taxpayer dollars, improve frontline worker safety, and deliver more reliable services to the community. https://lnkd.in/eGP3RGn9

Are you spending millions to build the city of tomorrow with yesterday's playbook? Very likely. Most urban planning is based on static models and siloed data. You approve a new high-rise, and suddenly traffic is gridlocked for 10 blocks. A new public transport policy looks great on paper but fails in the real world. These are multi-million dollar mistakes based on blind spots. What if you had your DT? It’s not just a 3D model. It's a breathing copy of your city, fed with real-time data. The Digital Twin City. It connects the physical to the digital, creating a dynamic feedback loop. With a city-scale Digital Twin, you can: ☑️Simulate Climate Impact: Model a 100-year flood and see exactly which assets are vulnerable before the storm hits. ☑️De-risk Construction: Visualize the ripple effects of new construction before breaking ground. ☑️Test Policies in Advance: See the unintended consequences of a new policy and refine it in the virtual world, not on your citizens. ☑️Unify Data: Break down silos between departments, sharing key data across jurisdictions to make holistic, informed decisions. ☑️Drive Investment: Clearly showcase development opportunities and the performance of city assets to attract investment. A DTC is the ultimate "try before you build" environment for creating smarter, more resilient cities. --------- Follow me for #digitaltwins Links in my profile Florian Huemer

🚴♀️ Cities worldwide constantly ask us for better insights into local cycling trends. Understanding active transport is vital for building healthier and more sustainable urban environments. That is why Google Earth has just launched a new "Cycling trip percentage" data layer. This tool gives cities and citizens access to neighbourhood level insights regarding active transport. Until now, gathering these baseline metrics has been highly demanding of time and resources. Today, mobility planners can instantly view local mobility patterns and compare their cycling rates with other cities across the globe. By making this information easily accessible, we can better support urban planning that is informed by solid evidence and create more sustainable cities for everyone. Learn more about this new data layer here: https://lnkd.in/et42McKR

The United Nations just released a new report on GeoAI ... “AI for Spatial Mapping and Analysis: GeoAI Toolkit for Urban Planners.” It explores how AI combined with geospatial data can support urban planning, infrastructure, climate resilience, and public health. Our team's paper was cited in this report "Generative AI for geospatial analysis: Fine-tuning ChatGPT to convert natural language into Python-based geospatial computations" Paper: https://lnkd.in/eZTh3fyK United Nations Report: https://lnkd.in/ergkbTHH UN-Habitat (United Nations Human Settlements Programme)