Physics: Universal patterns in the way people move

The frequency of trips made by people and distance travelled within and around cities follows a predictable and universal pattern, according to an analysis published this week in Nature. The research confirms the intuitive idea that people are unlikely to travel far too often. Predicting how people move within their cities and around the world is important for various areas of research such as urban planning and the modelling of epidemics.

The movement of people is fundamental to our societies, but a precise and quantitative description of human mobility has remained incomplete. Existing models, such as the gravity law or radiation model, concentrate on the spatial dependence of mobility flow, which is often insufficient to fully reproduce real-world data.

Lei Dong and colleagues analysed large-scale mobility data, based on anonymized mobile phone datasets, from several cities across the globe. These data were from Singapore, Greater Boston (USA), Dakar (Senegal), Abidjan (Ivory Coast) and several cities in Portugal, and were collected over different periods between 2006 and 2013. By taking into account not only the distance travelled but also the visitation frequency, they found that the number of individuals who visited different locations was highly consistent across the different cities. They also found that the number of visitors decreases in a predictable pattern for all locations in a given city, with respect to the frequency of visits and distance travelled. On this basis, the authors formulated a universal law of mobility, which enabled them to model and reproduce the temporal and spatial reach of population movements.

In an associated News & Views article, Laura Alessandretti and Sune Lehmann write that this study “identified a key component that was missing from existing theoretical frameworks of human mobility” — that is, the visitation frequency. They add that this finding, valid for various urban systems, “provides a general framework for describing and predicting mobility flows across timescales”.