Simulation of daily patterns of commuting and social activity

Principal Investigator: Mark Birkin (TALISMAN)
Co-investigators: David Martin (Hub), Andew Hudson-Smith (TALISMAN)

The TALISMAN node has an emerging interest in the development of new research methods for the exploitation of crowd-sourced data, in particular mobile telephone text messages and ‘tweets’ in the estimation of individual movement patterns in cities. We are especially interested in the deployment of modelling and simulation tools to predict daily patterns of commuting and intra-urban mobility for other purposes (such as shopping, leisure, education etc).

Members of the Hub team have complementary interests in the assessment of varying population ‘capacities’ at a variety of geo-located destinations such as schools, workplaces and retail centres. At the present time, relatively unsophisticated spatial re-weighting techniques have been used to reassign population to these destinations from associated sources of residential population, but much work has already been done to collate and standardise the handling of relevant data sources. The existing Population 24/7 work deals exclusively with aggregate patterns and not with individual behaviour.

These two streams are therefore highly complementary. The next step is to add greater definition in relation to trip destinations. The project team aims to produce new methods for the simulation of ‘social networks’ in UK cities showing how individual people move around and are brought together within different residential and non-residential environments. The aim is to produce a simulation which maps individuals not just by their place of usual residence (as in the Census for example) but traces their movement patterns around the city in small parcels of elapsed time (ie one hour or less).

The initial technical challenge is to define appropriate representations for activity patterns and their spatial manifestation. This will introduce distinctions between domestic life, work, education and retail/leisure. Places of residence will be defined at a neighbourhood scale of output area or less, while other activities will be resolved to discrete points in space (ie schools, retail parks etc). Census 2011 data will be used, and the project team also intend to exploit the value of other sources including pupil surveys, commercial market research, and location and movement patterns from mobile telephone devices.

There are many applications of micro-simulation in which the purpose is to reweight sample datasets to global populations and the project team therefore envisage that these methods will provide a suitable basis for reweighting social media despite their potential bias towards particular demographic and social groups.

The investigators are aware of the sensitivities in using individual data and at an early stage in the project a protocol will be established to govern the capture, storage, sharing, analysis and publication of data relating to individuals. The investigators' primary interest is in general activity patterns and behavioural rules so suitable safeguards can easily be established without compromising the research goals of the project.

Listen to David Martin talk about simulation of daily patterns of commuting and social activity in NCRM podcast series.