How often and from what distance do people come to certain places in cities? Researchers have investigated this question using mobile phone movement data from various metropolitan areas around the world – and in the process confirmed a universal rule: According to this, people visit places closer to home more often, while for special places they accept a longer journey, but get there less often. This connection can be described by an astonishingly simple mathematical formula. You can now help to model movement patterns, for example to optimize urban infrastructure or to predict the spread of new pandemics.
The patterns according to which people move is fundamental for our societies. Human mobility shapes the structure of cities, enables social exchange, causes traffic jams and environmental pollution and accelerates the spread of infectious diseases. There are many reasons for moving from one place to another: We go to work or shopping, visit friends, go on trips to parks and museums or go on vacation. How and how often we do this is relevant information for urban planners, but also for epidemiologists, for example, who want to predict and limit the spread of a pandemic.
Mobile phone movement data from four continents
A team led by Markus Schläpfer from the Massachusetts Institute of Technology (MIT) in Cambridge has now developed a new model that can be used to predict people’s mobility patterns. To do this, they evaluated anonymized mobile phone movement data from various metropolitan areas around the world, including the greater Boston area in the USA, Lisbon in Europe, Singapore in Asia and Dakar in Africa. In total, more than eight billion location-based data from more than four million people flowed into the study. “What we found out is that there is a very clear inverse relationship between the distance and the frequency of visits,” says Schläpfer’s colleague Paolo Santi. “You rarely go to far away places and you usually tend to go to places close by more often. That tells us how we organize our life. “
These results correspond to the intuitive assumption, but have never been proven in this form using a reliable database. “We may shop every day in a bakery a few hundred meters away, but we only go to the chic boutique that is miles from where we live once a month. This intuitive notion had never been tested empirically. When we did that, we found an incredibly regular and robust law – what we called the Visiting Law, ”says co-author Carlo Ratti.
Relationship between distance and frequency of visits
In mathematical terms, the visiting law says that the number of visitors to any urban location changes with the inverse square of both the distance from their place of residence and the frequency of visits. People who had to travel a shorter distance come more often than those who have to travel a long way. “Imagine you are in a busy square, say Boston, and you see people come and go. That might look pretty random and chaotic, but the law shows that these movements are surprisingly structured and predictable, ”explains Schläpfer. “Basically, it says how many of these people come from one, two or ten kilometers away and how many visit once, twice or ten times a month. And the best part is that this regularity not only applies in Boston, but in all cities around the world. “
The visiting law thus makes it possible to predict the flow of visitors between locations with an accuracy not previously achieved. Based on these predictions, urban planners can, for example, find the optimal place for a new shopping center, plan where attractive offers such as restaurants or parks should be set up, or optimize local public transport. The model can also be helpful for epidemiologists, for example in the event of a pandemic to predict where there is a particularly high risk of infection – and to initiate countermeasures if necessary.
Basis for improved models
Previous models had also focused on the distances covered, but ignored the frequency of visits. By including this, the authors have identified “a key component that was missing from the existing theoretical framework of human mobility,” said Laura Alessandretti and Sune Lehmann from the Technical University of Denmark in a comment on the study, also published in the journal Nature has been.
Alessandretti and Lehmann point out, however, that the current model still has weaknesses: On the one hand, it only allows predictions for urban areas. Whether there are similar patterns in rural regions with weaker infrastructure remains an open question. In addition, Schläpfer and his colleagues assumed for the analysis that the place of residence is the starting point for all movements. “But in the real world, geographic considerations and the need to minimize travel time mean that travel occurs in certain recurring sequences. People often visit places in a certain order, for example from work to the grocery store and the gym and then home, ”the comment said. Future models should therefore also consider such more complex movement patterns. From the point of view of Alessandretti and Lehmann, the current study paves the way for this.
Source: Markus Schläpfer (Massachusetts Institute of Technology, Cambridge) et al., Nature, doi: 10.1038 / s41586-021-03480-9