Learning from urban form to predict building heights

Understanding cities as complex systems, sustainable urban planning depends on reliable high-resolution data, for example of the building stock to upscale region-wide retrofit policies. For some cities and regions, these data exist in detailed 3D models based on real-world measurements. However, they are still expensive to build and maintain, a significant challenge, especially for small and medium-sized cities that are home to the majority of the European population. New methods are needed to estimate relevant building stock characteristics reliably and cost-effectively. Here, we present a machine learning based method for predicting building heights, which is based only on open-access geospatial data on urban form, such as building footprints and street networks. The method allows to predict building heights for regions where no dedicated 3D models exist currently. We train our model using building data from four European countries (France, Italy, the Netherlands, and Germany) and find that the morphology of the urban fabric surrounding a given building is highly predictive of the height of the building. A test on the German state of Brandenburg shows that our model predicts building heights with an average error well below the typical floor height (about 2.5 m), without having access to training data from Germany. Furthermore, we show that even a small amount of local height data obtained by citizens substantially improves the prediction accuracy. Our results illustrate the possibility of predicting missing data on urban infrastructure; they also underline the value of open government data and volunteered geographic information for scientific applications, such as contextual but scalable strategies to mitigate climate change.

[1]  S. Hellweg,et al.  Machine learning based modeling of households: A regionalized bottom‐up approach to investigate consumption‐induced environmental impacts , 2019, Journal of Industrial Ecology.

[2]  Linda See,et al.  City-descriptive input data for urban climate models: Model requirements, data sources and challenges , 2020 .

[3]  Hugh J W Sturrock,et al.  Predicting residential structures from open source remotely enumerated data using machine learning , 2018, PloS one.

[4]  Qi Zhou,et al.  Exploring the relationship between density and completeness of urban building data in OpenStreetMap for quality estimation , 2018, Int. J. Geogr. Inf. Sci..

[5]  Filip Biljecki,et al.  Generating 3D city models without elevation data , 2017, Comput. Environ. Urban Syst..

[6]  Christoph F. Reinhart,et al.  Urban building energy modeling – A review of a nascent field , 2015 .

[7]  Martin Fleischmann,et al.  momepy: Urban Morphology Measuring Toolkit , 2019, J. Open Source Softw..

[8]  Tianqi Chen,et al.  XGBoost: A Scalable Tree Boosting System , 2016, KDD.

[9]  Thomas Nauss,et al.  Importance of spatial predictor variable selection in machine learning applications - Moving from data reproduction to spatial prediction , 2019, Ecological Modelling.

[10]  P. Patel,et al.  Global scenarios of urban density and its impacts on building energy use through 2050 , 2017, Proceedings of the National Academy of Sciences.

[11]  Adam Millard-Ball,et al.  The world’s user-generated road map is more than 80% complete , 2017, PloS one.

[12]  Elsa Arcaute,et al.  Urban Science: Integrated Theory from the First Cities to Sustainable Metropolises , 2020, SSRN Electronic Journal.

[13]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[14]  Stefan Lüdtke,et al.  Flood loss estimation using 3D city models and remote sensing data , 2018, Environ. Model. Softw..

[15]  Pascal Neis,et al.  Quality assessment for building footprints data on OpenStreetMap , 2014, Int. J. Geogr. Inf. Sci..

[16]  Zhe Zhu,et al.  Understanding an urbanizing planet: Strategic directions for remote sensing , 2019, Remote Sensing of Environment.

[17]  Yoshua Bengio,et al.  Random Search for Hyper-Parameter Optimization , 2012, J. Mach. Learn. Res..

[18]  Filip Biljecki,et al.  Estimating Building Age with 3D GIS , 2017 .

[19]  D. Rybski,et al.  Settlement percolation: A study of building connectivity and poles of inaccessibility , 2019, Landscape and Urban Planning.

[20]  Muntaha Sakeena,et al.  Automatic Prediction of Building Age from Photographs , 2018, ICMR.

[21]  M. Barthelemy,et al.  A typology of street patterns , 2014, Journal of The Royal Society Interface.

[22]  Corinne Le Quéré,et al.  Urban infrastructure choices structure climate solutions , 2016 .

[23]  Magesh Nagarajan,et al.  A geospatial approach of downscaling urban energy consumption density in mega-city Dhaka, Bangladesh , 2018, Urban Climate.

[24]  Aliyu Salisu Barau,et al.  Six research priorities for cities and climate change , 2018, Nature.

[25]  Isabel M. Horta,et al.  A scenario-based approach for assessing the energy performance of urban development pathways , 2018, Sustainable Cities and Society.

[26]  Yu Liu,et al.  Spatial interpolation using conditional generative adversarial neural networks , 2019, Int. J. Geogr. Inf. Sci..

[27]  William F. Lamb,et al.  Upscaling urban data science for global climate solutions , 2019, Global Sustainability.

[28]  Geoff Boeing,et al.  OSMnx: New Methods for Acquiring, Constructing, Analyzing, and Visualizing Complex Street Networks , 2016, Comput. Environ. Urban Syst..

[29]  Filip Biljecki,et al.  Applications of 3D City Models: State of the Art Review , 2015, ISPRS Int. J. Geo Inf..

[30]  Isabel M. Horta,et al.  A spatially-explicit methodological framework based on neural networks to assess the effect of urban form on energy demand , 2017 .

[31]  Vítor Leal,et al.  Urban Form and Energy Demand , 2017 .

[32]  Christopher Tull,et al.  A data-driven predictive model of city-scale energy use in buildings , 2017 .

[33]  Gaël Varoquaux,et al.  Scikit-learn: Machine Learning in Python , 2011, J. Mach. Learn. Res..

[34]  Rishee K. Jain,et al.  Data-driven Urban Energy Simulation (DUE-S): A framework for integrating engineering simulation and machine learning methods in a multi-scale urban energy modeling workflow , 2018, Applied Energy.

[35]  Maria A. Brovelli,et al.  A New Method for the Assessment of Spatial Accuracy and Completeness of OpenStreetMap Building Footprints , 2018, ISPRS Int. J. Geo Inf..