An inherent limiting factor of human mobility prediction

2012

Abstract Much work has been done on predicting where is one going to be in the immediate future, typically within the next hour. By contrast, we address the open problem of predicting human mobility far into the future, a scale of months and years.

The massive amounts of geolocation data collected from mobile phone records has sparked an ongoing effort to understand and predict the mobility patterns of human beings. In this work, we study the extent to which social phenomena are reflected in mobile phone data, focusing in particular in the cases of urban commute and major sports events. We illustrate how these events are reflected in the data, and show how information about the events can be used to improve predictability in a simple model for a mobile phone user's location.

Nature Communications

Social structures influence human behavior, including their movement patterns. Indeed, latent information about an individual’s movement can be present in the mobility patterns of both acquaintances and strangers. We develop a “colocation” network to distinguish the mobility patterns of an ego’s social ties from those not socially connected to the ego but who arrive at a location at a similar time as the ego. Using entropic measures, we analyze and bound the predictive information of an individual’s mobility pattern and its flow to both types of ties. While the former generically provide more information, replacing up to 94% of an ego’s predictability, significant information is also present in the aggregation of unknown colocators, that contain up to 85% of an ego’s predictive information. Such information flow raises privacy concerns: individuals sharing data via mobile applications may be providing actionable information on themselves as well as others whose data are absent.

Expert Systems, 2020

Many aspects of our lives are associated with places and the activities we perform on a daily basis. Most of them are recurrent and demand displacement of the individual between regular places like going to work, school or other important personal locations. To accomplish these recurrent daily activities, people tend to follow regular paths with similar temporal and spatial characteristics, especially because humans are frequently looking for uniformity to support their decisions and make their actions easier or even automatic. In this work, we propose a method for discovering common pathways across users' habits from human mobility data. By using a density-based clustering algorithm, we identify the most preferable locations the users visit, we apply a Gaussian mixture model over these places to automatically separate among all traces, the trajectories that follow patterns in order to discover the representations of individual's habits. By using the longest common sub-sequence algorithm, we search for the trajectories that are more similar over the set of users' habits trips by considering the distance that pairs of users or habits share on the same path. The proposed method is evaluated over two real-world GPS datasets and the results show that the approach is able to detect the most important places in a user's life, detect the routine activities and identify common routes between users that have similar habits paving the way for research techniques in carpooling, recommendation and prediction systems.

Understanding human mobility patterns is crucial to fields such as urban mobility and mobile network planning. For this purpose, we make use of large-scale datasets recording individuals spatio-temporal locations, from eight major world cities: Beijing, Tokyo, New York, Paris, San Francisco, London, Moscow and Mexico City. Our contributions are twofold: first, we show significant similarities in people's mobility habits regardless of the city and nature of the dataset. Second, we unveil three persistent traits present in an individual's urban mobility: repetitiveness, preference for shortest-paths, and confinement. These characteristics uncover people's tendency to revisit few favorite venues using the shortest-path available.

PLoS ONE, 2014

Fundamental laws governing human mobility have many important applications such as forecasting and controlling epidemics or optimizing transportation systems. These mobility patterns, studied in the context of out of home activity during travel or social interactions with observations recorded from cell phone use or diffusion of money, suggest that in extra-personal space humans follow a high degree of temporal and spatial regularity -most often in the form of timeindependent universal scaling laws. Here we show that mobility patterns of older individuals in their home also show a high degree of predictability and regularity, although in a different way than has been reported for out-of-home mobility. Studying a data set of almost 15 million observations from 19 adults spanning up to 5 years of unobtrusive longitudinal home activity monitoring, we find that in-home mobility is not well represented by a universal scaling law, but that significant structure (predictability and regularity) is uncovered when explicitly accounting for contextual data in a model of in-home mobility. These results suggest that human mobility in personal space is highly stereotyped, and that monitoring discontinuities in routine room-level mobility patterns may provide an opportunity to predict individual human health and functional status or detect adverse events and trends.

Scientific Reports, 2013

In this study we analyze the travel patterns of 500,000 individuals in Cote d'Ivoire using mobile phone call data records. By measuring the uncertainties of movements using entropy, considering both the frequencies and temporal correlations of individual trajectories, we find that the theoretical maximum predictability is as high as 88%. To verify whether such a theoretical limit can be approached, we implement a series of Markov chain (MC) based models to predict the actual locations visited by each user. Results show that MC models can produce a prediction accuracy of 87% for stationary trajectories and 95% for non-stationary trajectories. Our findings indicate that human mobility is highly dependent on historical behaviors, and that the maximum predictability is not only a fundamental theoretical limit for potential predictive power, but also an approachable target for actual prediction accuracy. S tudies of mobility patterns and predictions of individual mobility trajectories are important in many research fields, such as mobile computing, epidemic modeling, traffic planning and disaster response 1-3 . Real-time locations visited by individuals are typically collected through mobile devices equipped with global-positioning system (GPS) capability, mobile phone cell towers, or wireless local area network (WLAN) access points.

Annals of Telecommunications, 2020

Human mobility patterns are associated with many aspects of our life. With the increase of the popularity and pervasiveness of smartphones and portable devices, the Internet of Things (IoT) is turning into a permanent part of our daily routines. Positioning technologies that serve these devices such as the cellular antenna (GSM networks), global navigation satellite systems (GPS), and more recently the WiFi positioning system (WPS) provide large amounts of spatio-temporal data in a continuous way (data streams). In order to understand human behavior, the detection of important places and the movements between these places is a fundamental task. That said, the proposal of this work is a method for discovering user habits over mobility data without any a priori or external knowledge. Our approach extends a density-based clustering method for spatio-temporal data to identify meaningful places the individuals' visit. On top of that, a Gaussian mixture model (GMM) is employed over movements between the visits to automatically separate the trajectories accordingly to their key identifiers that may help describe a habit. By regrouping trajectories that look alike by day of the week, length, and starting hour, we discover the individual's habits. The evaluation of the proposed method is made over three real-world datasets. One dataset contains high-density GPS data and the others use GSM mobile phone data with 15-min sampling rate and Google Location History data with a variable sampling rate. The results show that the proposed pipeline is suitable for this task as other habits rather than just going from home to work and vice versa were found. This method can be used for understanding person behavior and creating their profiles revealing a panorama of human mobility patterns from raw mobility data.

Multi-agent models for simulating the mobility behavior of the urban population are gaining momentum due to increasing computing power. Such models pose high demands in terms of input data in order to be reliably able to match real world behavior. To run the models a synthetic population mirroring typical mobility demand needs to be generated based on real world observations. Traditionally this is done using travel diary surveys, which are costly (and hence have relatively low sample size) and focus mainly on trip choice rather than on activities for an entire day. Thus in this setting the generation of synthetic populations either relies on resampling identical activity chains or on imposing independence of various trips occurring during the day. Both * Acknowledgments: This research was partially funded by the EMPORA project (financed by the Austrian KLIEN). Travel survey data has been made available from Verband Region Stuttgart which is gratefully acknowledged.

Transportation Research Procedia

Retrieving exhaustive information about individual mobility patterns is an essential step in order to implement effective mobility solutions. Despite their popularity, digital travel surveys still require a significant amount of inputs from the respondent. Consequently, they require great efforts from both respondents and analysts, and are limited to a relatively short period of timebetween a few weeks and a year. Driven by these motivations, the approach proposed in this paper uses mobile phone location history to automatically detect activity location without any interaction with the respondent. The proposed methodology uses raw location data together with a special indexing technique to calculate the probability of performing a certain activity in a certain location. It uses a heuristic rule to improve this estimation by considering the value of information over time. Finally, GIS data about the number of facilities located in a certain area is downloaded in real-time to further improve the overall estimation. Results of this exploratory study support the idea that the proposed approach can reconstruct complex mobility patterns while minimizing the number of active inputs from the respondent.