Using spatial autocorrelation and regression modelling, this project investigates whether road casualty risk is randomly distributed across London or reflects persistent structural inequalities — with implications for targeted Vision Zero intervention.
Using spatial data analysis, this report examines whether Airbnb is out of control in London, assessing rule-breaking, commercialisation, and neighbourhood impacts to evaluate proposed regulation targeting professional landlords.
Cycling is London’s greenest commute — but cyclists breathe its dirtiest air. This data report applies spatial quantitative methods to show that where cycling is most popular, pollution is worst, and cycle lanes alone can’t fix it.
This study reinterprets the SugarScape model as a firm-level market competition simulation, examining how competitive pressure, brand expansion, and scale efficiency shape firm survival and wealth inequality across static and dynamic markets.
This study builds an agent-based model of Midtown Manhattan to examine how the distinct vacant-cruising strategies of taxis and for-hire vehicles differentially contribute to urban traffic congestion, using TLC trip record data for spatial calibration.
A production-constrained Poisson spatial interaction model is calibrated to predict supermarket patronage across 775 Output Areas in the London Borough of Havering. Site A is recommended over Site B across all scenarios, remaining robust under oil price shocks and distance decay sensitivity analysis.
This study analyses the topological and flow-weighted resilience of the London Underground using degree, betweenness, and closeness centrality. Betweenness-based removal causes the most rapid network fragmentation. An alternative routing analysis identifies Embankment as the optimal substitute when Waterloo is closed.
If you have ever walked through the City of London, navigated the perfectly tiled blocks of Barcelona’s Eixample, or gotten hopelessly lost in the narrow alleys of Venice, you already have an intuition that these cities feel fundamentally different. But what happens when you ask a computer algorithm to read them?
Topic: Adapting the Level of Traffic Stress (LTS) Framework to a UK Context Case Study: West Midlands Area Supervisor: Prof. Duncan Smith (UCL CASA) Started: March 2026
Using spatial autocorrelation and regression modelling, this project investigates whether road casualty risk is randomly distributed across London or reflects persistent structural inequalities — with implications for targeted Vision Zero intervention.
Using spatial data analysis, this report examines whether Airbnb is out of control in London, assessing rule-breaking, commercialisation, and neighbourhood impacts to evaluate proposed regulation targeting professional landlords.
Cycling is London’s greenest commute — but cyclists breathe its dirtiest air. This data report applies spatial quantitative methods to show that where cycling is most popular, pollution is worst, and cycle lanes alone can’t fix it.
This study reinterprets the SugarScape model as a firm-level market competition simulation, examining how competitive pressure, brand expansion, and scale efficiency shape firm survival and wealth inequality across static and dynamic markets.
This study builds an agent-based model of Midtown Manhattan to examine how the distinct vacant-cruising strategies of taxis and for-hire vehicles differentially contribute to urban traffic congestion, using TLC trip record data for spatial calibration.
A production-constrained Poisson spatial interaction model is calibrated to predict supermarket patronage across 775 Output Areas in the London Borough of Havering. Site A is recommended over Site B across all scenarios, remaining robust under oil price shocks and distance decay sensitivity analysis.
This study analyses the topological and flow-weighted resilience of the London Underground using degree, betweenness, and closeness centrality. Betweenness-based removal causes the most rapid network fragmentation. An alternative routing analysis identifies Embankment as the optimal substitute when Waterloo is closed.