Research

My research follows 4 main axes:

Fostering the field of machine learning for urban sustainability

There are high expectactions about the potential role of machine learning for sustainable cities, but the field is still nascent. With colleagues, I have written several reviews on the state of the research on machine learning applications relevant to mitigating climate change in cities. A couple of others are in the making. I am also involved in education content creation at as Content Committee Chair at Climate Change AI, which includes for example a wiki, a course and tutorials.

Scaling data availability for sustainable urban planning

Even though more and more data on cities become available, there are still large data gaps that prevent potentially impactful research applications, for example comparative analyses of cities’ infrastructure. We have an ungoing project that aims to yield an inventory of existing open government data on buildings globally. I have developped a method to predict building heights from urban form, which we are currently extending to other building attributes including building ages and types, and other metrics relevant to urban studies. We are currently working on a project aiming to create EU-level database that gathers existing open data on buildings and fills the data gaps with machine learning.

Developing applied ML methods for analyzing urban form

Increasingly rich data on urban morphology, land use, built infrastructure, offer opportunities of new quantitative analysis of cities, as well as predictive tools. With colleagues, we are developping approaches using machine learning to find energy-relevant features of urban form that can inform sustainable urban planning. We started with the simplest ML methods and are progressively aiming to develop tailored spatial machine learning approaches. We are working in three direction: the links between buildings characteristics & urban form, urban mobility & urban form, and urban temperature & urban form.

Deploying decision tools to inform climate change mitigation in cities

The ultimate goal of this research is to deploy tools that can inform decisions towards more sustainable cities. I am currently focusing on tools for the buildings sector, which is a large contributor to global energy use and CO2 emissions. The lack of relevant granular data has constrained modelling efforts to either small-scale analysis, or coarse approaches. With colleagues, I am working on upscaling current city-scale models to the European leavel.