Narjes Abbasabadi, & Mehdi Ashayeri. (2025). From tweets to energy trends (TwEn2): social sensing–informed urban building energy modeling. Frontiers in Energy Research, 13. https://doi.org/10.3389/fenrg.2025.1688348
View Publication
Abstract
This study builds on our previous research, “From Tweets to Energy Trends (TwEn)” framework, which demonstrated the potential of social media interactions to inform urban energy predictions through a data-driven approach. Building on this foundation, we introduce TwEn2, a hybrid data-driven and physics-based modeling framework that advances the framework by enabling building-level analysis of urban-scale energy use with greater spatial specificity and enhanced applicability for urban building energy modeling (UBEM). The framework integrates geo-tagged social sensing data from the X platform with the U.S. Department of Energy’s prototype building models—focusing on mid-sized and large multifamily residential buildings across New York City’s boroughs—and benchmarking records. Tweet activity within building footprints is analyzed as a proxy for occupant presence and behavior, allowing assessment of correlations between human social dynamics and both measured and simulated monthly energy use across electricity, natural gas, and total energy consumption. By incorporating heigh fidelity social-sensing-data into physics-based simulations, TwEn2 improves predictive accuracy and enables UBEM informed by human activity patterns, validated against empirical benchmarking data. This framework provides a scalable, generalizable tool for urban energy modeling, planning, resilience and sustainability strategies, demonstrating the potential of social media as a real-time indicator of occupant dynamics to support informed energy management in cities.
Alam, M. S., & Abbasabadi, N. (2026). A CFD-Integrated Parametric Framework for Evaluating Passive Carbon-Capture Enclosure Performance. Architecture, 6(2), 65. https://doi.org/10.3390/architecture6020065
View Publication
Abstract
Integrating direct air carbon capture (DAC) into buildings offers a promising pathway for reducing atmospheric CO2, yet the role of architectural design in enhancing passive carbon-capture performance remains underexplored. This study presents a computational framework developed to optimize architectural design and enclosure geometry for enhanced passive airflow, using mass-flow rate as a proxy for the comparative assessment of carbon absorption potential. Implemented within Rhino3D and Grasshopper using Ladybug and Eddy3D, the workflow integrates weather data and CFD simulation to compute segmented mass-flow rates through stacked capture trays. The framework simplifies traditionally complex CFD processes by introducing a custom segmented mass-flow calculation approach that enables comparative performance assessment during early-stage design. Results confirm the validity of the proposed workflow, revealing that façade rotation can modify total mass flow by up to 96.5%; seasonal wind variability can cause airflow to range from approximately 8.5 kg/s in January to 169.5 kg/s in May in Seattle. Spatial configuration can alter airflow by up to an order of magnitude and introduce substantial spatial heterogeneity within capture zones. This research establishes a performance-driven design framework that enables architectural geometry to actively enhance passive carbon-capture integration, positioning building design as a measurable contributor to climate mitigation strategies. Ultimately, this work bridges architectural design and carbon-capture engineering, supporting interdisciplinary approaches to scalable, climate-responsive building systems.
Keywords
direct air capture; computational fluid dynamics; carbon-positive solutions; climate-responsive architecture; sustainable design; nature-based solutions; design computation
Abbasabadi, N., Moroseos, T. F., Ashayeri, M., & Meek, C. (2026). AI-Enhanced Urban Building Energy Modeling for Health-Driven Decarbonization in Vulnerable Communities. Architecture, 6(2), 84. https://doi.org/10.3390/architecture6020084
View Publication
Abstract
Retrofitting existing residential buildings is a critical strategy for achieving urban decarbonization while addressing public health disparities, particularly in communities disproportionately affected by environmental and socioeconomic stressors. This study presents a scalable urban building energy modeling framework that integrates physics-based simulations with machine learning to evaluate and prioritize health-driven retrofit strategies across residential building stocks. Synthetic datasets were generated through parametric simulations of representative building archetypes and retrofit scenarios, capturing variations in envelope performance, HVAC systems, infiltration rates, and ventilation strategies. Machine learning models were trained as surrogate predictors of building energy performance, enabling the rapid evaluation of retrofit impacts. A range of algorithms—including decision trees, random decision forests, gradient-boosting machines, support vector machines, k-nearest neighbors, and artificial neural networks—were evaluated. An artificial neural network implemented as a multilayer perceptron was selected for further analysis due to its strong predictive performance (R2 = 0.94) and ability to capture complex nonlinear relationships among retrofit variables. The final model used the Port optimization algorithm for stable convergence and improved generalization. The framework is applied to Seattle’s Duwamish Valley, a community experiencing disproportionate environmental and health burdens, and is generalizable and transferable to other cities with comparable residential building stocks across a range of climatic and environmental contexts. The results highlight retrofit priorities—particularly infiltration reduction, HVAC upgrades, and improved envelope performance—that deliver co-benefits for energy efficiency, indoor environmental quality, and occupant health. The results demonstrate that machine learning-enhanced physics-based UBEM can significantly accelerate retrofit evaluation while preserving the interpretability of simulation-based approaches. The proposed framework provides a scalable approach for identifying health-informed retrofit pathways that support equitable urban decarbonization.
Keywords
urban building energy modeling; machine learning; health-driven energy retrofits; indoor environmental quality; decarbonization
Worthy, A., Ashayeri Mehdi, Marshall, J. D., & Abbasabadi Narjes. (2026). A Framework for Augmenting Simulation-Based Building Energy Models with Earth Observational Microclimate Data Using Machine Learning Predictions. Urban Science, 10(7), 341. https://doi.org/10.3390/urbansci10070341
View Publication
Abstract
Accurate urban building energy modeling (UBEM) is constrained by mismatches between standard climate inputs and actual urban microclimate conditions. This study introduces a scalable, bottom-up, framework that integrates EnergyPlus building energy modeling simulation outputs with Earth observational and geographical-based urban morphology data, which are enhanced through machine learning techniques to improve energy demand predictions in urban settings. Applied to Los Angeles (LA), California, we evaluate the representativeness of typical meteorological year (TMYx) sampling sites against actual urban environmental conditions. We find that while satellite-derived surface temperatures show reasonable alignment with average city conditions, significant discrepancies are observed in urban form metrics such as tree cover, street cover, and building density, suggesting that TMYx stations should be placed in denser urban areas. We augment EnergyPlus simulations for 19 single-family buildings, with remote sensing data using machine learning models, to generate city-wide residential energy consumption heatmaps corrected for microclimate conditions. Models capture substantial intra-urban variation, with predicted energy use differing by approximately 10% between neighborhoods. Feature importance analysis highlights land surface temperature as a key predictor, underscoring its relevance to building energy research. We also find the majority of TMY3 sampling sites to be in low-vulnerability areas, underscoring the structural mismatch that is embedded in urban form and climate. This framework offers a scalable path for integrating urban microclimate effects into energy modeling to enable more precise and equitable energy policy and planning.
Keywords
urban microclimates; urban building energy use; Earth observational data; infrastructure equity
Abbasabadi, N., Worthy, A., Zhang, Z., Zhao, Y., & Liu, Y. (2026). An AI-augmented urban building energy Modeling framework integrating physics-based simulation, machine learning, and computer vision: The CityEnerTwin UBEM module. Energy and Buildings, 369, Article 117984. https://doi.org/10.1016/j.enbuild.2026.117984
View Publication
Abstract
Urban Building Energy Modeling (UBEM) plays a critical role in evaluating urban decarbonization strategies, but its accuracy and scalability are often constrained by simplified building archetype, limited façade-level data, and the computational cost of large-scale scenario analysis. This study presents the CityEnerTwin UBEM module, an artificial intelligence (AI)-augmented, physics-based framework that integrates modular workflows for computer vision–based façade semantic enrichment, localized archetype development, predictive and surrogate machine learning (ML), and climate-informed analysis to support scalable and computationally efficient urban energy modeling. Specifically, these workflows combine validated Vision Transformer (ViT)-based façade feature extraction methods for deriving material properties and window-to-wall ratio (WWR) from street-view imagery, K-means clustering for localized archetype development, and two complementary ML modeling streams: predictive models trained on measured operational energy data and surrogate models trained on synthetic UBEM simulation data for rapid retrofit scenario evaluation. Demonstrated on the University of Washington campus, the framework reduced campus-scale simulation error relative to measured energy consumption from 27.8% to 15.8%, while image-derived WWR inputs improved agreement for representative buildings from 23% to 20%. Predictive modeling using Random Forest (RF) achieved a test R2 of 0.88 and identified occupancy, building scale, and geometric efficiency as the dominant predictors of energy demand. Surrogate modeling using an Artificial Neural Network (ANN) achieved a test R2 of 0.98 and identified roof insulation, WWR, and shading depth as the most influential retrofit parameters, enabling rapid retrofit scenario evaluation. The results demonstrate the potential of integrated physics-based and data-driven workflows to improve UBEM representation, scalability, and computational efficiency for urban decarbonization planning. As the analytical core of the broader CityEnerTwin platform, the module connects spatially indexed building data and model outputs with interactive visualization to support urban-scale energy analysis and decision making.
Keywords
Urban building energy Modeling; Physics-based simulation; Computer vision; Machine learning; Surrogate Modeling; Retrofit scenario analysis
Assistant Professor of Architecture Narjes Abbasabadi has been awarded one of 36 seed grants from AI@UW. Narjes’ project entitled “SEED-AI 2026: DesignAI – An AI-Augmented Framework for Design Learning” was selected for funding alongside a range of other projects. The project summary is included below. More information is available here. Artificial intelligence (AI) offers powerful new ways to support design reasoning, exploration, and data-informed decision-making, yet these capabilities remain largely untapped in architectural education. This project addresses a critical dual-literacy…
Liu, Y., & Abbasabadi, N. (2025). Enhancing urban building energy models with Vision Transformers: A Case study in material classification from Google street view. Energy and Buildings, 333, Article 115457. https://doi.org/10.1016/j.enbuild.2025.115457.
View Publication
Abstract
The growing urbanization and increased urban energy consumption highlight the need for energy use and greenhouse gas emissions reduction strategies. Urban Building Energy Modeling (UBEM) emerged as a valuable tool for managing and optimizing energy consumption at the neighborhood and city scales to support carbon reduction goals. However, the accuracy of the UBEM is often limited by the lack of large-scale building façade material dataset. This study introduces a new approach to enhance UBEM by integrating an automatic deep learning material classification pipeline. The pipeline leverages multiple views of Google Street View Images (SVIs) to extract building façade material information, utilizing two Swin Vision Transformer (ViT) models to capture both global and local features from the SVIs. The pipeline achieved a main material classification accuracy reached 97.08%, and the sub-category accuracy reached 91.56% in a multi-class classification task. As the first study to apply a deep learning model for material classification to enhance the UBEM framework, this work was tested on the University of Washington campus, which features diverse facade materials. The model demonstrated its effectiveness by achieving an overall accuracy increase of 11.4% in year-round total operational energy simulations. The scalability of this material classification pipeline enables a more accurate and cost-effective application of UBEM at broader urban scales.
Abbasabadi, N., & Ashayeri, M. (2024). Machine Learning in Urban Building Energy Modeling. In Abbasabadi, N., & Ashayeri, M. (Eds.), Artificial Intelligence in Performance-Driven Design : Theories, Methods, and Tools: Theories, Methods, and Tools. Wiley-Blackwell.
View Publication
Abstract
Urban building energy modeling (UBEM) plays a pivotal role in effective urban energy management and the holistic understanding of citywide energy performance. This book chapter delves into the integration of machine learning (ML) in UBEM, covering applications such as predictive energy consumption modeling and optimization, and providing insights into how ML techniques enhance modeling accuracy and efficiency. It explores current UBEM methods, highlighting their strengths and limitations, and discusses the opportunities presented by ML for advancing UBEM approaches. The chapter also introduces a hybrid UBEM approach that combines data-driven and physics-based simulations to enhance modeling accuracy and reduce uncertainties in capturing urban energy use. This fusion of ML and UBEM offers promising prospects for improving urban energy management practices.
Abbasabadi, N., & Ashayeri, M. (2024). Understanding Social Dynamics in Urban Building and Transportation Energy Behavior. In Abbasabadi, N., & Ashayeri, M. (Eds.), Artificial Intelligence in Performance-Driven Design : Theories, Methods, and Tools: Theories, Methods, and Tools. Wiley-Blackwell.
View Publication
Abstract
This chapter explores the impact of human dynamics, social determinants of public health ( SDPH ), mobility, and occupancy on urban energy use behavior, a topic previously overlooked due to individual buildings or transportation models. A novel, data-driven urban energy model is developed using Artificial Neural Networks ( ANN ), augmented by Garson, Lek's profile and Partial dependence Plot ( PDP ) methods, to holistically evaluate urban energy behavior across Chicago communities, integrating both building and transportation energy use. Utilizing diverse public datasets from the city of Chicago, and validated through cross-validation, the model assesses human dynamics in development of an integrated urban energy modeling. The findings reveal a significant association between SDPH status, mobility, occupancy, and urban energy behavior with household income being a major contributor post accounting for urban spatial patterns and building physical attributes. The study suggests that meeting decarbonization targets in cities requires a broader evaluation encompassing various urban energy determinants. It advocates for emerging technologies and detailed analytical scrutiny, urging researchers and policymakers towards a comprehensive understanding of urban energy use behaviors.
Abbasabadi, N., & Ashayeri, M. (2024). A Hybrid Physics-Based Machine Learning Approach for Integrated Energy and Exposure Modeling. In Abbasabadi, N., & Ashayeri, M. (Eds.), Artificial Intelligence in Performance-Driven Design : Theories, Methods, and Tools: Theories, Methods, and Tools. Wiley-Blackwell.
View Publication
Abstract
This chapter introduces a hybrid framework that brings machine learning (ML) and urban big data analytics into integrated modeling of indoor air quality, building operational energy, and ambient airflow dynamics. This holistic approach allows for more effective and accurate simulation results for the design of built environments that prioritize both climate and health considerations. To validate this framework, we undertook a pilot study on a naturally ventilated, large-size office building prototype, as provided by the U.S. Department of Energy. This prototype was hypothetically placed in a densely populated area of Downtown Chicago, IL. For our computations, we employed tools, including EnergyPlus, CONTAM, CFD0, and artificial neural networks (ANNs). The findings highlighted the proposed framework's robust ability to evaluate the effects of building energy efficiency strategies, such as natural ventilation. Additionally, it took into account the indoor concentration of outdoor pollution resulting from the implementation of such strategies. Employing the hybrid approach, we achieved an accuracy characterized by an R -squared value of up to 0.96, facilitated by ANNs. Compared to conventional physics-based simulation methods, the hybrid approach further accelerated the simulation process by up to 200 times. This novel framework offers valuable insights to architects and engineers during early-stage design decisions, enabling them to harmonize occupant health considerations with energy conservation objectives, thereby placing health and well-being at the forefront of decarbonization goals.