Urban data mining, intelligence and simulation

Scientific project leaders: Roberto Castello, Kavan Javanroodi
PhD student: Alina Walch
External: Nahid Mohajeri (University College London)

Factsheet 2019/2020

The group focuses on understanding urban systems through their physical processes, the modelling of their dynamics and of the energy potential in order to improve their environmental sustainability. Multiple approaches and techniques are used: (i) statistical analysis of urban data collected in situ or secondary datasets from clouds and national databases, (ii) usage of data driven and machine learning techniques for forecasting, classification and automated recognition of urban elements, (iii) simulation of environmental phenomena and energy fluxes using advanced modelling tools.

More specifically, urban systems, generally consisting of several building blocks, are analysed either by simulation or by measurement to account for the numerous interactions happening between the elementary building objects and their environment. These interactions can be radiative (with the exchange of shortwave and longwave), conductive and convective (through the exchange of heat) but can also relate to the impact of built surfaces on the air flow and trapping of heat in urban areas. The group also works on the integration of decentralized energy systems in urban areas by looking at the energy flows.

Due to the complexity of urban areas, machine-learning techniques and simulation tools are developed and used to maintain a balance between accuracy and computational time. The different monitoring campaigns provide useful data to validate and improve the models. Our research focuses on modelling the potential of renewable technologies (solar, wind and geothermal) in the built environment in relation to the variable nature of energy resources. Using the existing large amount of data from local communities and at the national/regional level, complemented by satellite imagery and remote sensing, we aim at performing accurate spatio-temporal assessment of renewable energy potential. For this, we combine state-of-the-art deterministic models, machine Learning techniques and data-driven methods and we use them to model the uncertainty associated to the predictions and to forecast long-term evolutions of the urban systems.

Research interests

Data and methodologies


  • Measurement of Turbulence in an Urban Setup (MoTUS)
  • Data collection campaigns
  • Satellite imagery analysis (MeteoSuisse)
  • Environmental and socio-economic secondary datasets
  • National and local surveys

Machine learning techniques

  • Estimate and forecast of RE potential at high spatial and temporal resolution
  • Classification and clustering of roofs, building types, district archetypes
  • Fast and efficient modelling of wind speed and stochastic phenomena in the built environment
  • Prediction uncertainty on forecasted values
  • Supervised and semi-supervised neural networks for satellite/aerial image processing

Deterministic modeling

  • Climate change and Urban climate
  • Energy demand
  • Outdoor thermal comfort
  • Energy systems

Current projects


Past research projects

See publications by former topic leaders Nahid Mohajeri / Jérôme Kämpf / Darren Robinson

  • The complexity of roof-shape and solar energy
  • Street canyon and solar energy
  • Data Mining: Geo-Dependent Energy Supply in Relation to Urban Form
  • Urban Configuration (Building Typology and Mobility Patterns) and Renewable Energy
  • Urban Metabolism: From Energy Flow to Energy Management
  • Integration of decentralized energy adaptive systems for cites (IDEAS4cities)
  • Innovative planning and management instruments of urban energy systems (QUAD – sustainable districts)
  • Sustainable cities and urban energy systems of the future: Urban multiscale energy modelling (UMEM)
  • Investigation of strategies leading to a 2000W city using bottom-up models of urban energy flows (2 kW)
  • Ecosystemic modelling of urban metabolism based on modern thermodynamics
  • Multiscale modelling of building-urban interactions: understanding, modelling and mitigating the heat island effect

PhD theses

9014 2021 Alina Walch Spatio-Temporal Estimation of Renewable Energy Potential in Built Environments using Big Data
9376 2019 Dan Assouline Machine Learning and Geographic Information Systems for Large-Scale Mapping of Renewable Energy Potential
7756 2017 Silvia Coccolo Bioclimatic Design of Sustainable Campuses using Advanced Optimisation Methods
6102 2014 Diane Perez A framework to model and simulate the disaggregated energy flows supplying buildings in urban areas
5673 2013 Urs Wilke Probabilistic Bottom-Up Modelling of Occupancy and Activities to Predict Electricity Demand in Residential Buildings >> Detailed thesis results (for access password please contact [email protected])
4587 2010 Frédéric Haldi On the unification of behavioural modelling, human comfort and energy simulation in buildings
4657 2010 Marylène Montavon Optimisation of Urban Form by the Evaluation of the Solar Potential
4548 2009 Jérôme Kämpf On the modelling and optimisation of urban energy fluxes
4531 2009 Adil Rasheed Multiscale modelling of urban climate



Covid-19 mobility restrictions: impacts on urban air quality and health

N. Mohajeri; A. Walch; A. Gudmundsson; C. Heaviside; S. Askari et al. 

Buildings & Cities. 2021-01-01. Vol. 2, num. 1, p. 759-778. DOI : 10.5334/bc.124.

Shallow geothermal energy potential for heating and cooling of buildings with regeneration under climate change scenarios

A. Walch; X. Li; J. Chambers; N. Mohajeri; S. Yilmaz et al. 

Energy. 2022-04-01. Vol. 244, p. 123086. DOI : 10.1016/j.energy.2021.123086.

Large-scale evaluation of the suitability of buildings for photovoltaic integration: Case study in Greater Geneva

M. Thebault; G. Desthieux; R. Castello; L. Berrah 

Applied Energy. 2022-06-15. Vol. 316, p. 119127. DOI : 10.1016/j.apenergy.2022.119127.

Impact of the COVID-19 pandemic on the energy performance of residential neighborhoods and their occupancy behavior

V. Todeschi; K. Javanroodi; R. Castello; N. Mohajeri; G. Mutani et al. 

Sustainable Cities And Society. 2022-07-01. Vol. 82, p. 103896. DOI : 10.1016/j.scs.2022.103896.

Quantifying the technical geothermal potential from shallow borehole heat exchangers at regional scale

A. Walch; N. Mohajeri; A. Gudmundsson; J-L. Scartezzini 

Renewable Energy. 2020. Vol. 165, p. 369-380. DOI : 10.1016/j.renene.2020.11.019.

Spatio-Temporal Relationship between Land Cover and Land Surface Temperature in Urban Areas: A Case Study in Geneva and Paris

X. Ge; D. Mauree; R. Castello; J-L. Scartezzini 

ISPRS International Journal of Geo-Information. 2020. Vol. 9, num. 10, p. 1-25, 593. DOI : 10.3390/ijgi9100593.

Big data mining for the estimation of hourly rooftop photovoltaic potential and its uncertainty

A. Walch; R. Castello; N. Mohajeri; J-L. Scartezzini 

Applied Energy. 2020-01-28. Vol. 262, p. 114404. DOI : 10.1016/j.apenergy.2019.114404.

Solar cooking potential in Switzerland: Nodal modelling and optimization

T. Chatelain; D. Mauree; S. Taylor; O. Bouvard; J. Fleury et al. 

Solar Energy. 2019-12-01. Vol. 194, p. 788-803. DOI : 10.1016/j.solener.2019.10.071.

A machine learning approach for mapping the very shallow theoretical geothermal potential

D. Assouline; N. Mohajeri; A. Gudmundsson; J-L. Scartezzini 

Geothermal Energy. 2019-07-25. Vol. 7, num. 1, p. 19. DOI : 10.1186/s40517-019-0135-6.

Integrating urban form and distributed energy systems: Assessment of sustainable development scenarios for a Swiss village to 2050

N. Mohajeri; A. T. D. Perera; S. Coccolo; L. Mosca; M. Le Guen et al. 

Renewable Energy. 2019-12-01. Vol. 143, p. 810-826. DOI : 10.1016/j.renene.2019.05.033.

Publications under Vahid Nik
Publications under Jérôme Kaempf
Publications under Darren Robinson