A place where data, AI and behavioural theory come together
Cities are hotspots of human activity that have increased our prosperity, happiness, and health. Yet the future liveability of cities around the world is under pressure as they face major social-technical challenges. Those challenges include crumbling social cohesion, income inequality, overcrowding of public spaces, and unhealthy local environments caused by factors such as heavy traffic and noise pollution. CityAI Lab examines the pivotal role that the urban environment plays in tackling such challenges.
Our research focusses on unravelling how the urban environment and human behaviour dance a tango. It capitalises on advances in machine learning and on the wealth of data available now at a city level. We combine these with established theories on planning and behaviour, hoping to contribute to the development of more attractive and liveable cities.
The CityAI Lab is part of the TU Delft AI Labs programme.
Slots for supervision of MSc students usually fill up rapidly, so be quick! We are always looking for highly motivated students who want to work with us on the intersections between (travel) behaviour modelling, machine learning, and urban (mobility) analytics.
The ideal student for an MSc project at CityAI Lab has:
- Knowledge of machine learning
- Knowledge of the domain of application, such as travel behaviour, transport system, etc.
- Relevant programming skills (e.g. Python, R, Matlab)
If you are interested in starting an MSc project on these topics, please send a brief motivation summary email to Sander, Oded, or Simeon, clarifying:
- What topic you are interested in
- Intended starting date
- Your relevant experiences (projects, courses, etc.)
- Programming skills (languages)
Examples of ongoing and completed MSc & BSc Projects
- Explaining urban space perceptions (Ruben Sangers)
- Measuring the Evolution of Social Segregation using Public Transport Smart Card Data (Lukas Kolkowski)
- TULIPS (European Green Deals): Sustainable Inter-modal transport connections, using data driven approaches.
- Explainable AI: A Proof of Concept Demonstration in Financial Transaction Fraud Detection using TreeSHAP & Diverse Counterfactuals (Pratheep Balakrishnan)
- Blending discrete choice modelling and computer vision (Joris van Eekeren)
- Bus Management using Multi-agent Reinforcement Learning (George Weijs).
- Automated Disruption Detections in Metro Networks using Smart Card Data (Faye Jasperse)