Are you a Returner or an Explorer? Ask Big Data

You are here

Are you a Returner or an Explorer? Ask Big Data

What were the first “gifts” Christopher Columbus and the other European explorers brought to indigenous people when they landed in the Americas? Don’t try to guess it: epidemics. Spanish invaders conquered an entire continent with very little effort, with invisible weapons unknown to the indigenous people: smallpox, influenza, chicken pox, measles. “One means by which Europeans were able to spread at the expense of other peoples”, writes American scientist and writer Jared Diamond, “was by infecting them (usually unintentionally) with epidemic infectious diseases […], to which Europeans had evolved some genetic resistance […] while unexposed non-European peoples had no such exposure, hence no such resistance”. The rapidity of the contagion and mortality rates were strikingly high: in about fifty years between 80% and 95% of the indigenous population of the Americas died. European explorers opened a new world, diffusing European culture, languages and opening new commercial routes. However, they never moved alone, because their diseases went with them in the exploration of new places, cities, continents.

What about today? Who are the Columbus and Vespucci’s in this modern, technological world where traveling is no longer an activity reserved for daring Renaissance men? Could such modern “explorers” be disease diffusers in their long, adventurous trips?

These are some of the fascinating questions my colleagues at KDD Lab and I investigated during a thrilling collaboration with Filippo Simini (University of Bristol) and Albert-László Barabási (Northeastern University in Boston and CEU University in Budapest). And to be coherent with our research questions we acted as explorers ourselves, distributing and developing the research in different places both in the Old and New Worlds: Pisa, Budapest and Boston. During three exciting years I split my life between these three amazing cities, investigating the patterns of individual mobility by analyzing phone calls and car displacements of about 100,000 individuals.

Big Data and Human Mobility

The availability of Big Data on human mobility allowed us to observe in detail the displacements on the territory of about 100,000 individuals. Two datasets were available for our study. The first one stored all the calls made by 50,000 (anonymized) mobile phone users during three months. Every time a user makes a call, indeed, the career stores his position allowing us to reconstruct her mobility trajectory. The second dataset stores information about more than 10 million trips made, during one month, in the region of Tuscany by 50,000 private vehicles with on-board GPS devices. Every time a person turns on the vehicle, a GPS installed into the car automatically starts sending data to a server with striking accuracy. This allowed us to  draw the “nervous system” of mobility, a picture that well outlines the complexity of our mobile ecosystem.


A fragment of the GPS trajectories used in our study, displaying trips originating in the metropolitan areas of Pisa (in blue) and Florence (red). This plain geo-referenced visualization of experimental data reveals the confrontation of two ‘competing’ metropolitan areas. It also demonstrates the ability of Big Data to portray social complexity.
A fragment of the GPS trajectories used in our study, displaying trips originating in the metropolitan areas of Pisa (in blue) and Florence (red). This plain geo-referenced visualization of experimental data reveals the confrontation of two ‘competing’ metropolitan areas. It also demonstrates the ability of Big Data to portray social complexity. Original figure here.

Starting from the data, we synthesized each individual’s mobility by computing her mobility radius, that is the characteristic distance traveled by the individual during the three months. A small mobility radius means that the individual typically travels within short distances, a high mobility radius indicates an individual covering large distances every day. Comparing the recurrent mobility radius of an individual (computed on her routine displacements, like home-work displacements) with her total mobility radius (computed on all her displacements) we discovered an interesting and unexpected result: people naturally split into two well-separated groups, with rather different mobility characteristics.

Returners and Explorers

The first group consists of people for whom the recurrent mobility radius is very similar to the total mobility radius. In other words the whole mobility of returners, as we named individuals in this mobility profile, can be reduced to the mere displacements they do between a few preferred locations (home and work places for example). The other group of individuals consists instead of explorers, people whose recurrent mobility is just a small fraction of their overall mobility. Explorers do not reduce their mobility to a predefined home-work routine, presenting a high degree of exploration. Their mobility radius is hence defined by the displacements between all the locations they visit.


forked_plot
The total mobility radius of individuals (x axis) versus their recurrent mobility radius (y axis). We observe that the individuals concentrated around the diagonal (returners) have a total mobility radius comparable to their recurrent mobility radius. The individuals concentrated around the abscissa (explorers) have a recurrent mobility radius considerably smaller than total mobility radius.

You are a Columbus or not, hence, no middle ground. And if you are a Columbus, your mobility has its own peculiar shape. In contrast to returners, who show a bi-modal mobility between home and work places, explorers show a star-like mobility: a central core of locations (composed by home and work places) around which distant cores of locations gravitate (see Figure below). These locations are the “new lands” an explorer visits in her trips, when visiting something new between Pisa and Florence or somewhere in the beautiful Tuscany landscapes. They are somewhat like Columbus, being capable of destroying the everyday routine to explore new “Americas”, new satellites to add to their main mobility core.


thumbnail
The mobility networks of returners and explorers. Nodes (circles) indicate the geographic locations visited by the individual, and each link denotes a travel observed between two locations. When the total mobility radius (rg) is small, the two most important locations (home and work places, red and blue) are close to each other for both explorers and returners. As the total radius increases the behaviour of returners and explorers starts to differ; for returners, the two most important locations move away from each other; for explorers, they stay close and other clusters of locations emerge far from the centre of mass (the grey cross).

Infector Explorers

As Columbus and his colleagues, if you are an explorer when you travel you are not alone. Yet, your diseases follow you diffusing in the new neighborhoods, cities or continents you visit. Our intensive experiments and computer simulations showed indeed that returners and explorers exhibit different properties in the spreading of epidemics on a territory. Explorers are indeed more dynamic actors: they diffuse new ideas faster but they are also responsible of disease spreading, because they cover a larger territory in a smaller time than returners. On this aspect, therefore, the explorers of today are not so different from European sailors: they use airplanes and cars instead of caravels, but they can be as dangerous as Renaissance explorers were.

But the surprises are not over, because our analyses also found that marriages are made in heaven: by observing the social network of individuals inferred from mobile phone data, we discovered that explorers preferably communicate with other explorers and returners preferably communicate with other returners. The way you explore your mobility sphere, hence, also determines your social network.

“So you are an explorer, after so much traveling between Pisa, Budapest and Boston!”, László Barabási told me when I showed him the existence of the returners/explorers dichotomy. “You are an explorer too, László, I think all scientists are explorers”, I replied. Some categories of people are definitely explorers. Scientists, who move relentlessly around the world as part of their job, lie in this category. From conferences to project meetings and collaborations, scientists diffuse knowledge and ideas across the world: their mobility is characterized by a plenty of distant satellites gravitating around the home-university mobility core. Other people, conversely, are definitely returners. High school students, for example, are characterized by a home-school routine, presenting more likely what we call a bi-modal mobility.

What about you? Are you a returner or an explorer? Likely, you already know the answer without recurring to our mathematical model. But our research tells you something more about the structure of your mobility, and your role in society. If you are like Columbus you have a star-like mobility, and you communicate with other globe trotters. If you are a returner maybe your mobility can be less adventurous, but at least the next flu epidemic will be not your fault.


This post refers to the following publication:

L. Pappalardo, F. Simini, S. Rinzivillo, D. Pedreschi, F. Giannotti, A.-L. Barabasi, Returners and Explorers dichotomy in Human Mobility, Nature Communications, 6:8166, doi: 10.1038/ncomms9166, 2015.

You can find more material and download the paper here and here.


Related projects: