Detecting and comparing localised structure in economic and spatial networks

Project ID: 2228cd1257 (You will need this ID for your application)

UCL Lead department: Centre for Advanced Spatial Analysis (CASA)

Project Summary:

From social networks to transport networks, protein networks to economic networks, networks are ubiquitous in daily life. While there are a wide range of tools to probe the structure and function of networks, one of the most well-known and widely applied is community detection. Community detection algorithms uncover clusters of tightly nodes, and are often based on dynamical processes. For social and economic networks, these processes often correspond to the diffusion of knowledge, opinions and ideas. The presence of clusters or communities in these networks ‘trap’ this knowledge and constrain its diffusion. While the literature on community detection is deep and ever growing, there remain significant open questions that emerge once communities have been detected.

One of the key areas of interest here is that of modular network comparison. Despite the vast range of potential applications, existing approaches compare communities as sets and thus neglect the internal structure. While some progress has been made on ‘global’ network comparison by our research group (Straulino et al 2021, EPJ), this project focuses on comparing ‘local’ community structure. For example, in a social network with two types of ties – social media and classroom ties – some friendship groups might be very similar in both networks, but others less so.

While these methods are general to most network types, we will apply them to economic and spatial networks with a focus on industry and occupation labour flow networks capturing the flow of knowledge, technological networks quantifying the recombinative processes underlying innovation, and mobility networks that unveil the spatial structure of cities. There will be an opportunity to link to local and national policy through existing links (e.g., DLUHC).

This project would suit a candidate with a strong quantitative background (maths, physics, engineering, computer science) and an interest in economics, economic geography or cities.