One metaphor for movement of people within a city is that of blood flowing through a body. We speak of the ‘pulse of the city,’ for example. Can we assess a city’s health by monitoring the flow of people, just like a nurse takes your heart-rate and blood pressure during a health check? Can we find neighbourhoods in need of help by looking at the flow patterns of the public transport system?
Using data from automated fare collection in the London rail system, we trained a classifier on several features, such as:
- Passenger flow. The normal expected flow between neighbourhoods can be modeled like gravity, proportional to the population of the areas and inversely proportional to the distance between them. Greater or lesser flow than expected is a clue to the socio-economic health of a neighbourhood.
- Diversity of connections. A neighbourhood with frequent trips to a large number of other places may have more economic activity, and thus be healthier.
- Transport modality. Poorer areas rely more on buses, rather than trains.
Using these features and others, our classifier could identify areas of high deprivation as measured by the Indices of Multiple Deprivation, including poverty, unemployment, housing problems, crime, etc. On the composite index, our model has precision 0.805, sensitivity 0.733, and specificity 0.810.
Although the results are limited by the available information, we have demonstrated a strong potential for a real-time neighbourhood health monitor – a city nurse – which could support city planners, local authorities, and citizens who want to:
- Get early warning about areas of high deprivation
- Make more efficient choices about regeneration and renewal
- Assess the effects of regeneration projects
- Hold local authorities to account