Periodic Reporting for period 1 - UDENE (Urban Development Explorations using Natural Experiments)
Reporting period: 2024-01-01 to 2024-12-31
Urbanization is increasing globally, with almost all cities growing in population This growth is more prominent in Europe and North Africa, where 70% and 78% of people, respectively, live in cities. Climate change, inequalities and conflict drive huge populations to cities, in search of a better life. However, urban development is not generally guided by evidence-based informed decisions, especially in emerging countries. As a result, liveability in cities worsen. The UDENE Project aims to use large amounts of Earth Observation (EO) data from Copernicus satellites and local sources to support evidence-based decision making for urban development. To support the UN Sustainable Development Goal 11 of making cities safe, resilient, and sustainable, the project will essentially create a virtual laboratory for urban planners and visionaries to test their development ideas. Multi-dimensional models of urban areas across time and across countries will be used to support urban planners in exploring natural occurrences of similar development ideas, concepts, or projects in similar urban areas (natural experiments). The wealth of EO data in fast accessible Data Cube format provides a great search space for exploring natural experiments, by exploiting causal analysis. Project outcomes will be embodied by an exploration tool for natural experiments and a match-making tool which will leverage the exploration tool by linking the existing EO product, process, or service offerings from the EU. Project activities include Open (FSTP) Calls.
During the first reporting period of the UDENE project, significant advancements were made in data integration, causal inference models, and tool development to support evidence-based urban planning. The project successfully built and operationalized a Data Cube to structure and process in-situ and Earth Observation data. Multi-source data collection was conducted, incorporating environmental, urban, and socioeconomic datasets. The Data Cube was linked to the Copernicus Data Space Ecosystem and external EO data sources such as CreoDIAS, enabling global data access and integration.
The project developed hierarchical regression models and machine learning-based algorithms to analyze the causal effects of urban planning decisions on environmental and economic outcomes. Three use cases were implemented, each focusing on different urban challenges: in Serbia, assessing pollution and traffic impact due to new transportation infrastructure; in Tunisia, evaluating urban heat island effects and green infrastructure strategies; and in Türkiye, conducting earthquake preparedness analysis for high-rise districts. Sensitivity analysis was applied to validate the models, ensuring their robustness for decision-making.
A key milestone was the development and testing of the Exploration and Matchmaking Tools. The Exploration Tool was designed to allow urban planners to identify and compare urban development patterns based on natural experiments, while the Matchmaking Tool was integrated to connect planners from Copernicus Internation Partners with relevant EO products and services from the EU countries. Initial testing of these tools began in the fourth quarter of 2024, with further refinements and expanded dataset integration planned.
The main achievements of the project include the successful operationalization of the UDENE Data Cube, establishing a functional and scalable EO data infrastructure for urban development research. The integration of in-situ and historical datasets allows dynamic scenario analysis. Significant progress was made in advancing causal effect modeling, developing data-driven methodologies to identify and analyze natural experiments in urban planning, and applying sensitivity analysis to validate urban development interventions.
The deployment of decision-support tools marks a major milestone, with prototype versions of the Exploration Tool and Matchmaking Tool released, enabling city planners to conduct evidence-based analysis using real-world data. Additionally, international data integration and collaboration were enhanced by strengthening data-sharing agreements with Copernicus International Partners and facilitating cross-border collaborations to improve the Copernicus In-Situ Component datasets.
The project developed hierarchical regression models and machine learning-based algorithms to analyze the causal effects of urban planning decisions on environmental and economic outcomes. Three use cases were implemented, each focusing on different urban challenges: in Serbia, assessing pollution and traffic impact due to new transportation infrastructure; in Tunisia, evaluating urban heat island effects and green infrastructure strategies; and in Türkiye, conducting earthquake preparedness analysis for high-rise districts. Sensitivity analysis was applied to validate the models, ensuring their robustness for decision-making.
A key milestone was the development and testing of the Exploration and Matchmaking Tools. The Exploration Tool was designed to allow urban planners to identify and compare urban development patterns based on natural experiments, while the Matchmaking Tool was integrated to connect planners from Copernicus Internation Partners with relevant EO products and services from the EU countries. Initial testing of these tools began in the fourth quarter of 2024, with further refinements and expanded dataset integration planned.
The main achievements of the project include the successful operationalization of the UDENE Data Cube, establishing a functional and scalable EO data infrastructure for urban development research. The integration of in-situ and historical datasets allows dynamic scenario analysis. Significant progress was made in advancing causal effect modeling, developing data-driven methodologies to identify and analyze natural experiments in urban planning, and applying sensitivity analysis to validate urban development interventions.
The deployment of decision-support tools marks a major milestone, with prototype versions of the Exploration Tool and Matchmaking Tool released, enabling city planners to conduct evidence-based analysis using real-world data. Additionally, international data integration and collaboration were enhanced by strengthening data-sharing agreements with Copernicus International Partners and facilitating cross-border collaborations to improve the Copernicus In-Situ Component datasets.
UDENE Research Advancements over the State of the Art are as below:
Novi Sad Use Case
The satellite-based approaches offer broader environmental coverage and robust sensitivity analysis. There are methods that provide sophisticated handling of unmonitored regions and multi-factor urban analysis. We advance the field through transportation-specific modeling and explicit spatial relationship preservation. Our focus on explainability (PFI, IG, OAT) and urban planning applications fills crucial gaps in current research.
Tunis Use Case
Our approach surpasses state-of-the-art methods in data integration, modeling, and sensitivity analysis. Unlike prior studies using basic land cover data or climate simulations, we integrate six-channel EO data (NDVI, NDBI, NDWI, SAVI, EVI, IBI) with rigorous validation. Instead of standard models, we employ a U-Net architecture to capture both local and global heat patterns. Our sensitivity analysis goes beyond previous metrics, incorporating ablation studies, perturbation analysis, statistical testing, spatial occlusion, and gradient-based interpretability. These advancements enable precise spatial analysis, robust validation, and superior pattern recognition. Our future work will integrate UTCI analysis and urban morphology metrics to further refine urban heat assessments.
Istanbul Use Case
Our approach integrates detailed seismic, geological, and urban data with dual GMPEs for a two-step hazard and loss assessment—specifically tailored for high-rise buildings.
While SOTA works offer static parameter evaluation or future simulation-based uncertainty, our method delivers continuous loss monitoring and advanced sensitivity analysis. This framework not only deepens seismic analysis but also allows integration of complementary techniques, such as Fuzzy TOPSIS for location assessment and social impact considerations from Cremen’s model.
Novi Sad Use Case
The satellite-based approaches offer broader environmental coverage and robust sensitivity analysis. There are methods that provide sophisticated handling of unmonitored regions and multi-factor urban analysis. We advance the field through transportation-specific modeling and explicit spatial relationship preservation. Our focus on explainability (PFI, IG, OAT) and urban planning applications fills crucial gaps in current research.
Tunis Use Case
Our approach surpasses state-of-the-art methods in data integration, modeling, and sensitivity analysis. Unlike prior studies using basic land cover data or climate simulations, we integrate six-channel EO data (NDVI, NDBI, NDWI, SAVI, EVI, IBI) with rigorous validation. Instead of standard models, we employ a U-Net architecture to capture both local and global heat patterns. Our sensitivity analysis goes beyond previous metrics, incorporating ablation studies, perturbation analysis, statistical testing, spatial occlusion, and gradient-based interpretability. These advancements enable precise spatial analysis, robust validation, and superior pattern recognition. Our future work will integrate UTCI analysis and urban morphology metrics to further refine urban heat assessments.
Istanbul Use Case
Our approach integrates detailed seismic, geological, and urban data with dual GMPEs for a two-step hazard and loss assessment—specifically tailored for high-rise buildings.
While SOTA works offer static parameter evaluation or future simulation-based uncertainty, our method delivers continuous loss monitoring and advanced sensitivity analysis. This framework not only deepens seismic analysis but also allows integration of complementary techniques, such as Fuzzy TOPSIS for location assessment and social impact considerations from Cremen’s model.