Understanding individual mobility patterns from urban sensing data: A mobile phone trace example
Citations
635 citations
634 citations
Cites background from "Understanding individual mobility p..."
...Because cell phones are almost universally used, data about their use have led to important findings about movement in urban (Calabrese et al. 2010; Kang et al. 2012), regional (Calabrese et al. 2013; Sagl et al. 2012), and national scales (Krings et al. 2009; Simini et al. 2012)....
[...]
...2012), regional (Calabrese et al. 2013; Sagl et al. 2012), and national scales (Krings et al....
[...]
431 citations
425 citations
Cites background or methods or result from "Understanding individual mobility p..."
...The locations reported in CDR are cell phone tower locations and thus depend on the density of the cellular network, which varies from as little as a few hundred meters in metropolitan areas to a few kilometers in rural regions (e.g., Calabrese et al., 2013; Chen et al., 2014)....
[...]
...…in transportation research, the rapid rise and prevalence of mobile technologies have enabled the collection of a massive amount of passive data (big data), which have resulted in a surge of studies on human movement (e.g., Gonzalez et al., 2008; Kang et al., 2012a, 2012b; Calabrese et al., 2013)....
[...]
...62 using the prevailing approach in the literature (Calabrese et al., 2013)....
[...]
...However, since most of the studies have tried to mine activity locations instead of trajectories8 and few studies have attempted validating (Calabrese et al., 2013; Chen et al., 2014), the meaning of a displacement still is unclear....
[...]
...As noted earlier, the limited amount of work on validation is mostly at the area level, for example, comparing population density calculated from the big data against the census data (Calabrese et al., 2013), or comparing shares of trip purposes against a travel survey (Gong et al....
[...]
402 citations
References
5,514 citations
844 citations
762 citations
"Understanding individual mobility p..." refers result in this paper
...It should be noted that the associations found in this study do not necessarily mean causality since residential self-selection may confound the association between the built environment and travel behavior (see Mokhtarian and Cao, 2008)....
[...]
...During the last two decades, we have seen an explosion in the deployment of pervasive systems like cellular networks, GPS devices, and WiFi hotspots that allow us to capture massive amounts of real-time data related to people and cities (Reades et al., 2007; Gonzalez et al., 2008; Wang et al., 2009). The usage of these datasets could enable researchers to better understand the laws governing people’s movements and improve the efficiency and responsiveness of urban policies. This study aims to explore the potential value and challenges of these novel datasets in urban modeling, using the mobile phone trace data collected by mobile network operators as an example. Compared to travel survey data, the mobile phone trace data provide researchers new opportunities to examine individual mobility from an alternative perspective with their lower collection cost, larger sample size, higher update frequency, and broader spatial and temporal coverage. The mobile phone locations are routinely collected by operators for network management purposes, therefore, the datasets are theoretically available at no cost to researchers. The datasets allow for studying individual mobility of millions of people across the metropolitan area over a longer time period compared to a few thousand households’ movements within 1–2 days usually collected through travel surveys. They are updated on a real time basis, which could lead to more reliable and trackable urban performance indicators and support more prompt policy responses to emerging urban issues. Meanwhile, the mobile phone trace data also have significant drawbacks for transportation research: (1) socio-economic and demographic attributes are not available due to privacy concerns, which are indispensable to calibrate models at disaggregate level to explore the underlying behavior mechanism of individual/household mobility choice; (2) mobile phone users might not represent a random sample of the population. The results need careful analyses to be properly interpreted; and (3) the datasets are not primarily designed for modeling purposes and are often not in an easy-to-use format, which restricts the usefulness of raw data without intensive processing. This study represents a first step towards building a methodology to utilize mobile phone data for transportation research. In this study, we use mobile phone traces from about one million users in the Boston Metropolitan Areas, Massachusetts, USA over 3 months to characterize individual mobility and understand its spatial patterns within a metropolitan area. To address the lack of socio-economic and demographic attributes in mobile phone data, we aggregate mobility measures generated from individual mobile phone traces to block groups, the most disaggregate level of census geography at which socio-economic and demographic information is available, and associate the aggregate mobility measures with census data. Given the aggregate nature of our analysis, we raise two cautions at the outset. First, the underlying behavior mechanism cannot be identified by this study. The second caution concerns the ecological fallacy, which is the fallacy related to inferring the nature of individuals based solely upon aggregate statistics collected for the group. Therefore, it should be noted that what we find in this study is the general spatial patterns of mobility within the metropolitan area and their relationships to neighborhood characteristics, but not how individuals’ characteristics and built environment influence their own travel behavior. Nonetheless, using aggregate data collected for a long time period could help screen the idiosyncratic factors at the individual level, identify the underlying trends, and explain the variation in intra-urban mobility patterns. As Yang (2008) and Yang and Ferreira (2008) demonstrate, even without individual preference, urban spatial structure alone could explain a significant portion of the variation in commuting distance. Similar analytic approaches that involve data aggregation have been adopted by many previous studies in mobility research. Lindsey et al. (2011) explore the relationship between Vehicle Kilometers Traveled (VKT) and urban form characteristics at grid cell level. Yang (2008) examines the relationship between excess commuting distance and urban spatial structure at census tract level. Holtzclaw et al. (2002) find that the average annual distance driven per car at the Traffic Analysis Zone (TAZ) level is a strong function of density, income, household size and public transit....
[...]
...During the last two decades, we have seen an explosion in the deployment of pervasive systems like cellular networks, GPS devices, and WiFi hotspots that allow us to capture massive amounts of real-time data related to people and cities (Reades et al., 2007; Gonzalez et al., 2008; Wang et al., 2009). The usage of these datasets could enable researchers to better understand the laws governing people’s movements and improve the efficiency and responsiveness of urban policies. This study aims to explore the potential value and challenges of these novel datasets in urban modeling, using the mobile phone trace data collected by mobile network operators as an example. Compared to travel survey data, the mobile phone trace data provide researchers new opportunities to examine individual mobility from an alternative perspective with their lower collection cost, larger sample size, higher update frequency, and broader spatial and temporal coverage. The mobile phone locations are routinely collected by operators for network management purposes, therefore, the datasets are theoretically available at no cost to researchers. The datasets allow for studying individual mobility of millions of people across the metropolitan area over a longer time period compared to a few thousand households’ movements within 1–2 days usually collected through travel surveys. They are updated on a real time basis, which could lead to more reliable and trackable urban performance indicators and support more prompt policy responses to emerging urban issues. Meanwhile, the mobile phone trace data also have significant drawbacks for transportation research: (1) socio-economic and demographic attributes are not available due to privacy concerns, which are indispensable to calibrate models at disaggregate level to explore the underlying behavior mechanism of individual/household mobility choice; (2) mobile phone users might not represent a random sample of the population. The results need careful analyses to be properly interpreted; and (3) the datasets are not primarily designed for modeling purposes and are often not in an easy-to-use format, which restricts the usefulness of raw data without intensive processing. This study represents a first step towards building a methodology to utilize mobile phone data for transportation research. In this study, we use mobile phone traces from about one million users in the Boston Metropolitan Areas, Massachusetts, USA over 3 months to characterize individual mobility and understand its spatial patterns within a metropolitan area. To address the lack of socio-economic and demographic attributes in mobile phone data, we aggregate mobility measures generated from individual mobile phone traces to block groups, the most disaggregate level of census geography at which socio-economic and demographic information is available, and associate the aggregate mobility measures with census data. Given the aggregate nature of our analysis, we raise two cautions at the outset. First, the underlying behavior mechanism cannot be identified by this study. The second caution concerns the ecological fallacy, which is the fallacy related to inferring the nature of individuals based solely upon aggregate statistics collected for the group. Therefore, it should be noted that what we find in this study is the general spatial patterns of mobility within the metropolitan area and their relationships to neighborhood characteristics, but not how individuals’ characteristics and built environment influence their own travel behavior. Nonetheless, using aggregate data collected for a long time period could help screen the idiosyncratic factors at the individual level, identify the underlying trends, and explain the variation in intra-urban mobility patterns. As Yang (2008) and Yang and Ferreira (2008) demonstrate, even without individual preference, urban spatial structure alone could explain a significant portion of the variation in commuting distance. Similar analytic approaches that involve data aggregation have been adopted by many previous studies in mobility research. Lindsey et al. (2011) explore the relationship between Vehicle Kilometers Traveled (VKT) and urban form characteristics at grid cell level. Yang (2008) examines the relationship between excess commuting distance and urban spatial structure at census tract level. Holtzclaw et al. (2002) find that the average annual distance driven per car at the Traffic Analysis Zone (TAZ) level is a strong function of density, income, household size and public transit. Wang (2001) explains intraurban variations of commuting time and distance at TAZ level in Columbus, Ohio using Census Transportation Planning Package (CTPP) data....
[...]
...During the last two decades, we have seen an explosion in the deployment of pervasive systems like cellular networks, GPS devices, and WiFi hotspots that allow us to capture massive amounts of real-time data related to people and cities (Reades et al., 2007; Gonzalez et al., 2008; Wang et al., 2009). The usage of these datasets could enable researchers to better understand the laws governing people’s movements and improve the efficiency and responsiveness of urban policies. This study aims to explore the potential value and challenges of these novel datasets in urban modeling, using the mobile phone trace data collected by mobile network operators as an example. Compared to travel survey data, the mobile phone trace data provide researchers new opportunities to examine individual mobility from an alternative perspective with their lower collection cost, larger sample size, higher update frequency, and broader spatial and temporal coverage. The mobile phone locations are routinely collected by operators for network management purposes, therefore, the datasets are theoretically available at no cost to researchers. The datasets allow for studying individual mobility of millions of people across the metropolitan area over a longer time period compared to a few thousand households’ movements within 1–2 days usually collected through travel surveys. They are updated on a real time basis, which could lead to more reliable and trackable urban performance indicators and support more prompt policy responses to emerging urban issues. Meanwhile, the mobile phone trace data also have significant drawbacks for transportation research: (1) socio-economic and demographic attributes are not available due to privacy concerns, which are indispensable to calibrate models at disaggregate level to explore the underlying behavior mechanism of individual/household mobility choice; (2) mobile phone users might not represent a random sample of the population. The results need careful analyses to be properly interpreted; and (3) the datasets are not primarily designed for modeling purposes and are often not in an easy-to-use format, which restricts the usefulness of raw data without intensive processing. This study represents a first step towards building a methodology to utilize mobile phone data for transportation research. In this study, we use mobile phone traces from about one million users in the Boston Metropolitan Areas, Massachusetts, USA over 3 months to characterize individual mobility and understand its spatial patterns within a metropolitan area. To address the lack of socio-economic and demographic attributes in mobile phone data, we aggregate mobility measures generated from individual mobile phone traces to block groups, the most disaggregate level of census geography at which socio-economic and demographic information is available, and associate the aggregate mobility measures with census data. Given the aggregate nature of our analysis, we raise two cautions at the outset. First, the underlying behavior mechanism cannot be identified by this study. The second caution concerns the ecological fallacy, which is the fallacy related to inferring the nature of individuals based solely upon aggregate statistics collected for the group. Therefore, it should be noted that what we find in this study is the general spatial patterns of mobility within the metropolitan area and their relationships to neighborhood characteristics, but not how individuals’ characteristics and built environment influence their own travel behavior. Nonetheless, using aggregate data collected for a long time period could help screen the idiosyncratic factors at the individual level, identify the underlying trends, and explain the variation in intra-urban mobility patterns. As Yang (2008) and Yang and Ferreira (2008) demonstrate, even without individual preference, urban spatial structure alone could explain a significant portion of the variation in commuting distance. Similar analytic approaches that involve data aggregation have been adopted by many previous studies in mobility research. Lindsey et al. (2011) explore the relationship between Vehicle Kilometers Traveled (VKT) and urban form characteristics at grid cell level....
[...]
...During the last two decades, we have seen an explosion in the deployment of pervasive systems like cellular networks, GPS devices, and WiFi hotspots that allow us to capture massive amounts of real-time data related to people and cities (Reades et al., 2007; Gonzalez et al., 2008; Wang et al., 2009). The usage of these datasets could enable researchers to better understand the laws governing people’s movements and improve the efficiency and responsiveness of urban policies. This study aims to explore the potential value and challenges of these novel datasets in urban modeling, using the mobile phone trace data collected by mobile network operators as an example. Compared to travel survey data, the mobile phone trace data provide researchers new opportunities to examine individual mobility from an alternative perspective with their lower collection cost, larger sample size, higher update frequency, and broader spatial and temporal coverage. The mobile phone locations are routinely collected by operators for network management purposes, therefore, the datasets are theoretically available at no cost to researchers. The datasets allow for studying individual mobility of millions of people across the metropolitan area over a longer time period compared to a few thousand households’ movements within 1–2 days usually collected through travel surveys. They are updated on a real time basis, which could lead to more reliable and trackable urban performance indicators and support more prompt policy responses to emerging urban issues. Meanwhile, the mobile phone trace data also have significant drawbacks for transportation research: (1) socio-economic and demographic attributes are not available due to privacy concerns, which are indispensable to calibrate models at disaggregate level to explore the underlying behavior mechanism of individual/household mobility choice; (2) mobile phone users might not represent a random sample of the population. The results need careful analyses to be properly interpreted; and (3) the datasets are not primarily designed for modeling purposes and are often not in an easy-to-use format, which restricts the usefulness of raw data without intensive processing. This study represents a first step towards building a methodology to utilize mobile phone data for transportation research. In this study, we use mobile phone traces from about one million users in the Boston Metropolitan Areas, Massachusetts, USA over 3 months to characterize individual mobility and understand its spatial patterns within a metropolitan area. To address the lack of socio-economic and demographic attributes in mobile phone data, we aggregate mobility measures generated from individual mobile phone traces to block groups, the most disaggregate level of census geography at which socio-economic and demographic information is available, and associate the aggregate mobility measures with census data. Given the aggregate nature of our analysis, we raise two cautions at the outset. First, the underlying behavior mechanism cannot be identified by this study. The second caution concerns the ecological fallacy, which is the fallacy related to inferring the nature of individuals based solely upon aggregate statistics collected for the group. Therefore, it should be noted that what we find in this study is the general spatial patterns of mobility within the metropolitan area and their relationships to neighborhood characteristics, but not how individuals’ characteristics and built environment influence their own travel behavior. Nonetheless, using aggregate data collected for a long time period could help screen the idiosyncratic factors at the individual level, identify the underlying trends, and explain the variation in intra-urban mobility patterns. As Yang (2008) and Yang and Ferreira (2008) demonstrate, even without individual preference, urban spatial structure alone could explain a significant portion of the variation in commuting distance....
[...]
662 citations
589 citations
"Understanding individual mobility p..." refers background in this paper
...As a result, there are not many respondents included in any one neighborhood, which limits the efforts to adequately understand travel patterns for small areas (Handy, 1996)....
[...]