Urban Informatics Using Mobile Network Data

This introductory chapter presents the key concepts of urban informatics, discusses the use of mobile network data in urban informatics, and provides an outline of the book’s structure and scope. It discusses traditional methods in understanding travel behavior i.e., survey-based approach. A surveyed information is used as an input for travel demand models that predict future travel behavior and transportation needs from the current travel behavior data. A survey-based approach is however costly and time-consuming, and hence its information can be dated and limited to a particular region. Mobile network data thus emerges as a promising alternative data source that is massive in both cross-sectional and longitudinal perspectives, which provides a wider geographic coverage of travelers and a longer time of travel behavior observation. From the point of view of urban informatics, there are new opportunities as well as challenges in utilizing such data in travel behavior research. This chapter discusses the state of the art and outlines the rest of the book, which is a series of case studies done by our research teams based on mobile phone data collected from different countries to tackle different steps and approaches in travel demand modeling.

Urban transportation is a key issue worldwide, especially in Sub-Saharan Africa where there has been a rapid increase in population and at the same time, a lack of infrastructure which include railways, airways and roads. People’s mobility in these African nations is mostly provided by bus services and a large-scale informal public transportation system known as paratransit (for example, car rapides in Senegal, Tro Tros in Ghana, taxis in Uganda and Ethiopia, and Matatus in Kenya). This brings up the need for transport demand estimation, which is a challenging task, particularly in developing countries. The main reason for the challenge is that the estimation methods usually require large datasets which can be quite difficult, costly, and time-consuming to collect. When it comes to demand estimation, important factors include the accuracy and transparency of data. Accurate data can help us identify trends in passenger demand so that better informed decisions about future investments in infrastructure and capacity can be made. In this chapter, we discuss how passenger demand for public transportation services can be estimated using mobile phone network data. Based on the inferred travel demand, strategic locations for public transportation services such as paratransit and taxi stands can then be suggested accordingly. This chapter is inspired by our original research work done by Demissie et al. (IEEE Trans Intell Transp Syst. 2016;17(9):2466–78; 13th international conference on electrical engineering/electronics, computer, telecommunications and information technology (ECTI-CON). IEEE; 2016).

Creation of a trip distribution model necessitates a large amount of data collection, such as costly travel surveys to determine trip makers’ origin and destination zones. Intrazonal trips are frequently overlooked when developing trip distribution models due to the difficulties in calculating their travel expenses. Ignoring intra-zonal travels, on the other hand, leads to inaccurate model estimates, especially when the zone size is large and the number of intra-zonal visits is high. This is especially relevant given the increased interest worldwide in making cities more walkable and bikeable, where intra-zonal travel accounts for a large portion of those journeys. We employ mobile phone network data to generate country-wide mobility trends in this chapter. For 123-district level traffic analysis zones, a set of doubly constrained trip distribution models that integrate intrazonal trips is estimated. We then examine two methods for calculating intra-zonal travel expenses based on trip distance. The average intra-zonal trip distances determined from the two techniques yields differing levels of sensitivity to the distance-decay effect, according to our analysis. This demonstrates that in the absence of intrazonal trips, model estimation produces erroneous results. The content discussed in this chapter is inspired our original work by Demissie et al. (IEEE Trans Intell Transp Syst. 2019;20(7):2605–17; 13th international conference on electrical engineering/electronics, computer, telecommunications and information technology (ECTI-CON); 2016. p. 1–6).

Understanding the causes and impacts of migration, as well as implementing policies aimed at providing certain services, requires estimating migration flows and forecasting future patterns. Over time, less study has been done on modeling migration flows than has been done on modeling other types of flows, such as commutes. One of the biggest hurdles to empirical analysis and theoretical developments in the modeling of migration flows has been a lack of data. Because a migration trip is far less frequent than a commute, it necessitates a longitudinal set of data for study. The data from a large mobile phone network is used in this chapter to infer migration trips and their distribution. Intra/inter-district migration flows, migration distance distribution, and origin-destination (O-D) movements are among the interesting properties of the inferred migration trips. The log-linear model, classic gravity model, and recently developed radiation model are investigated for migration trip distribution modeling, with distinct approaches applied in setting parameters for each model. As a result, among the different models, gravity and log-linear models with a direct distance (displacement) as a trip cost and district centroids as reference points perform the best. Among the radiation models, a model that considers district population is the best performing model, but not as good as the gravity and log-linear models. This chapter reflects the idea and thinking process of our original work by Phithakkitnukoon et al. (IEEE Access. 2022;10:23248–58; IEEE international conference on privacy, security, risk and trust and IEEE international conference on social computing (PASSAT/SocialCom 2011); 2011. p. 515–20), and Hankaew et al. (IEEE Access. 2019;7(1):164746–58).

Previous chapters have shown how mobile network data can be used to infer travel generation and trip distribution. We continue to explore further in this chapter into how to utilize the mobile network data in making inferences about transport mode choice and its social influence. This chapter focuses on social influence in terms of ego-network effect in commuting mode choice, for which a longitudinal mobile phone data that includes both location and communication records is investigated. Methods for inferring social tie strength and transport mode as well as a framework for analyzing social influence in transport mode choice are discussed. The findings reveal that a person’s strong relationships are more essential in determining whether or not driving is the person’s preferred mode of transportation, whereas weak ties are more relevant in determining whether or not public transportation is the person’s preferred mode of transportation. The data also shows that social ties that are geographically nearby have a greater influence on commuting mode choice than those that are farther away. In the case of public transportation, accessibility distance is also a deciding factor. As the access distance increases, the percentage of people who use public transportation decreases. Furthermore, the social network has been found to influence commute mode choice, with the likelihood of choosing a given mode increasing as the percentage of social ties who choose that mode increases. The content discussed in this chapter reflects the idea, motivation, and thinking process in our original work done by Phithakkitnukoon et al. (EPJ Data Sci. 2017;6(11); Soc Netw Anal Min. 2016;6(1)).

This chapter describes a framework for using a mobile phone GPS data to investigate the effects of weather on people’s daily activity routines. Temperature, rainfall, and wind speed are among the weather variables considered in our case study discussed in this chapter. We describe a method for inferring people’s daily activity patterns, including the places they visit and when they do so, as well as the duration of the visit, based on GPS position traces of their mobile phones and Yellow Pages information. An analysis of 31,855 mobile phone users reveals that people are more likely to stay at restaurants or food outlets for longer periods of time, and to a lesser extent at retail or shopping sites, when the weather is extremely cold or the ambiance is calm (non-windy). People’s activity habits are affected by certain weather conditions when compared to their usual patterns. People’s motions and activities are evident at different times of the day. The weather has a wide range of effects on different geographical areas of a large city. When urban infrastructure data is employed to characterize areas, significant connections between weather conditions and people’s accessibility to public rail network are observed. This chapter gives a new perspective of how mobile phone GPS data can be utilized in the context of weather’s influence on human behavior, specifically choices of daily activities, as well as the impact of environmental factors on urban life dynamics. The conceptual framework and analysis discussed in this chapter are based on the original research by Phithakkitnukoon et al. (PLoS One. 8:12, 2013; PLoS One. 7:10, 2012; Activity-aware map: Identifying human daily activity pattern using mobile phone data. LNCS, 2010).

This chapter discusses a framework for analyzing tourist behavior that employs a large-scale opportunistic mobile sensing approach through a case study of mobile phone users in Japan. It deliberates on how enormous mobile phone GPS location records can be utilized in analyzing tourist travel behavior through a number of proxies, including the number of trips made, time spent at places, and mode of transportation used. Furthermore, this chapter examines the relationship between personal mobility and tourist travel behavior, and it reveals a number of meaningful insights for tourism management, including tourist flows, top tourist destinations and origins, top destination types, top modes of transportation in terms of time spent and distance traveled, and how personal mobility information can be used to estimate a likelihood in tourist travel behavior, such as the number of trips initiated, time spent at destination, and trip distance. In addition, the chapter describes an application developed based on the findings of this analysis allowing its user to monitor touristic, non-touristic, and commuting trips, as well as home and work locations and tourist flows, which can be useful for city planners, transportation managers, and tourism authorities. This chapter is inspired by our original work by (Phithakkitnukoon et al., Pervasive Mob Comput 18: 2015) and Horanont et al. (Lecture notes of the institute for computer sciences, social-informatics and telecommunications engineering, LNICST: 2015).

This chapter provides an outlook for future directions of mobile network data-based urban informatics, particularly in travel behavior research. It discusses a dynamic characteristic of mobile network data that continues to change its properties and potential values with the technological advancement, which in turn poses new challenges and exciting research opportunities. This may include new paradigms for data collection as an alternative to the telecom provided data, which is rarely accessible due to data privacy regulations. There is still a need for rebalancing between the data’s utility and privacy for which data uncertainty and privacy algorithms are discussed. Future directions of mobile network data-based urban informatics will concern data mining techniques that help discover patterns and trends in trajectory data, including group movement pattern mining, trajectory clustering, and sequential pattern mining. Mining such patterns benefits travel behavior research with many applications, such as traffic detection, social gathering recognition, regional travel behavioral signature extraction, uncovering life/daily patterns, and unusual event detection.