Hannah Gumble and Nicolas Palominos and I put together this poster for the 2022 GIS Research UK (GISRUK2022) conference. It came about through our shared interest in developing metrics of how street space is allocated and used in cities. It seems that the urgent need to shift travel...

What far-fetched transport tech will lure commentators to declare ‘peak mobility innovation’ this year? Perhaps it’ll be something like exercise bike company Peloton’s CommuCycle concept - an exercise bike in a car so you can cycle while you drive. It’s an extreme example of the proliferation of (generally more sensible)...

This paper was presented at the GeoSim workshop at SIGSPATIAL 2020. I also made a poster summarising the work.

Sequential sampling models are well validated in psychology and represent decisions resulting from gradual acumulation of information, typically at milisecond time scales. In this paper I develop a sequential sampling model for a pedestrian’s choice of where to cross the road and implement this in an agent-based model (ABM).

The nice thing about using a sequential sampling model in an ABM is that agents interact with one another as they gradually form choice preferences which allows interactions to impact agents’ choices in a comprehensive and flexible way. In this case, pedestrians choose where to cross the road based on local traffic conditions and the availability of crossing infrastructure.

Originally published in Local Transport Today, Jul 24, 2020 (pg. 22)

In case you hadn’t noticed, driverless cars are not going to be neatly delivering us about the city any time soon (despite claims to the contrary here, here and here). But while city streets respond...

This was a piece of analysis for the British Red Red Cross’ COVID-19 Vulnerability Index. The index brings togather a variety of small area open data that help to measure the vulnerability of the local population to COVID-19.

Open data on hospital locations in England and Wales were identified (see this notebook for a summary of these data sources) and geocoded using the ONS Postcode Directory. Then using the Ornance Survey Open Roads dataset and the speedy python network analysis library pandana the distances from LSOA centroid to the hospitials were calculated.

This python script contains all the details and the indicator data is here.

Thanks to Tom Russell and the British Red Cross Maps Team for their help.

I helped to organise the first CASA Doctoral Summer School. Bonnie Buyuklieva spearheaded the project with a successful funding bid to EPSRC for doctoral student projects related to the UK’s Industrial Strategy. Together with Matt Ng we put together a program of short data science methods presentations and workshops delivered by academics, mostly from CASA but also other departments and unis.

Each presentation was accompanied by a code notebook and you can access these here. The participants worked in groups to produce analysis related to the Industrial Strategy and you can see their work here.

CASA Phds Metteo Niera, Melisa Barrientos, Nicolas Palominous and I took part in a two day hackathon organised by the Freight Traffic Control 2050 project. Full poster here.

Using CitySprint delivery data we were given access to, we scoped out the potential to break up deliveries into two sections: deliveries by vehicles to small hubs and delivery by foot from hubs to final destinations. Our proposal won first prize!

(full disclosure: there were only two teams, the other team was also from CASA)