Fine-grained prediction of urban population using mobile phone location data

Chen, Jie; Pei, Tao; Shaw, Shih‐Lung; Lü, Feng; Li, Mingxiao; Cheng, Shifen; Gao, Liang; Zhang, Hengcai

doi:10.1080/13658816.2018.1460753

Cited by 56 publications

(45 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, for the situation of a citywide metro network, it is difficult to achieve high-quality feature engineering manually. The most important feature for the prediction of the citywide spatiotemporal passenger flow volume is the spatiotemporal dependencies [14,15]. In a metro network, the spatial dependencies refer to the interactions between the passenger inflow volume and outflow volume in near and distant stations.…”

Section: Introductionmentioning

confidence: 99%

Predicting Station-Level Short-Term Passenger Flow in a Citywide Metro Network Using Spatiotemporal Graph Convolutional Neural Networks

Han

Wang

Ren

et al. 2019

IJGI

View full text Add to dashboard Cite

Predicting the passenger flow of metro networks is of great importance for traffic management and public safety. However, such predictions are very challenging, as passenger flow is affected by complex spatial dependencies (nearby and distant) and temporal dependencies (recent and periodic). In this paper, we propose a novel deep-learning-based approach, named STGCNNmetro (spatiotemporal graph convolutional neural networks for metro), to collectively predict two types of passenger flow volumes-inflow and outflow-in each metro station of a city. Specifically, instead of representing metro stations by grids and employing conventional convolutional neural networks (CNNs) to capture spatiotemporal dependencies, STGCNNmetro transforms the city metro network to a graph and makes predictions using graph convolutional neural networks (GCNNs). First, we apply stereogram graph convolution operations to seamlessly capture the irregular spatiotemporal dependencies along the metro network. Second, a deep structure composed of GCNNs is constructed to capture the distant spatiotemporal dependencies at the citywide level. Finally, we integrate three temporal patterns (recent, daily, and weekly) and fuse the spatiotemporal dependencies captured from these patterns to form the final prediction values. The STGCNNmetro model is an end-to-end framework which can accept raw passenger flow-volume data, automatically capture the effective features of the citywide metro network, and output predictions. We test this model by predicting the short-term passenger flow volume in the citywide metro network of Shanghai, China. Experiments show that the STGCNNmetro model outperforms seven well-known baseline models (LSVR, PCA-kNN, NMF-kNN, Bayesian, MLR, M-CNN, and LSTM). We additionally explore the sensitivity of the model to its parameters and discuss the distribution of prediction errors.

show abstract

Section: Introductionmentioning

confidence: 99%

Predicting Station-Level Short-Term Passenger Flow in a Citywide Metro Network Using Spatiotemporal Graph Convolutional Neural Networks

Han

Wang

Ren

et al. 2019

IJGI

View full text Add to dashboard Cite

show abstract

“…For example, Figure is a product of tracking activity diary data of over 500 households (J. Chen et al, ). The concept is equally applicable to tracing mediated forms of location and mobility as illustrated in studies of up to 17 million residents of Shanghai (e.g., B. Y. Chen et al, ; J. Chen, Pei, et al, ; B. Y. Chen, Wang, et al, ).…”

Section: Introductionmentioning

confidence: 99%

“…The concept is equally applicable to tracing mediated forms of location and mobility as illustrated in studies of up to 17 million residents of Shanghai (e.g., B. Y. Chen et al, 2016;J. Chen, Pei, et al, 2018;B.…”

Section: Introductionmentioning

confidence: 99%

Trace of pace, place, and space in personal relationships: The chronogeometrics of studying relationships at scale

Spitzberg

2019

Personal Relationships

View full text Add to dashboard Cite

The study of personal relationships has traditionally relied on self‐reports or observations of face‐to‐face interaction. Digital media increasingly provide the ability to trace communication and relationships at scale. Such methods portend significant theoretical and methodological challenges, as well as potential. As a way of illustrating such potential, big data approaches to the select traditional relational concepts of routine relating, propinquity, homophily, small world, and reciprocity are reviewed. The fields of communication and personal relationships will need to inform such research by developing their own interdisciplinary relationships with geographic information sciences, computational linguistics, and computer sciences or cede a significant frontier of their field to these other disciplines.

show abstract

“…This situation makes it possible to obtain large amounts of dynamic spatiotemporal data with precise spatial information [17,18]. Among the various types of urban big data, mobile phone location data, which can represent approximately the entire urban population, are the most widely used big data in obtaining fine-grained population distributions [19][20][21][22][23][24][25] and conducting dynamic population mapping [26,27].…”

Section: Introductionmentioning

confidence: 99%

Population Distributions of Age Groups and Their Influencing Factors Based on Mobile Phone Location Data: A Case Study of Beijing, China

et al. 2019

Self Cite

View full text Add to dashboard Cite

The fine-grained population distributions of different age groups are crucial for urban planning applications. With the development of information and communication technology (ICT), detailed population data retrieved from various big data sources, especially on a fine scale, have been extensively used for urban planning. However, studies estimating the detailed population distributions of different age groups are still lacking. This study constructs a framework to generate fine-grained population data for different age groups and explores the influence of various factors on the distributions of different age groups. The population is divided into the following four age groups: (1) early adulthood people: 18 ≤ age ≤ 24, (2) young people: 25 ≤ age ≤ 39, (3) middle-aged people: 40 ≤ age ≤ 59, and (4) elderly people: 60 ≤ age. The results indicate that education and accommodation factors have a major influence on the distributions of early adulthood and elderly people, respectively. Business, restaurant, and accommodation factors are the main factors influencing the population distributions of young and middle-aged people. The accommodation factor plays a major controlling role at night, and its explanatory power gradually decreases during the day, while the explanatory powers of the business and restaurant factors increase and become leading factors during the day. Specifically, the hospital factor has a greater effect on the distribution of elderly people. The entertainment factor has very little explanatory power for the population distributions of the different age groups.

show abstract

Fine-grained prediction of urban population using mobile phone location data

Cited by 56 publications

References 18 publications

Predicting Station-Level Short-Term Passenger Flow in a Citywide Metro Network Using Spatiotemporal Graph Convolutional Neural Networks

Predicting Station-Level Short-Term Passenger Flow in a Citywide Metro Network Using Spatiotemporal Graph Convolutional Neural Networks

Trace of pace, place, and space in personal relationships: The chronogeometrics of studying relationships at scale

Population Distributions of Age Groups and Their Influencing Factors Based on Mobile Phone Location Data: A Case Study of Beijing, China

Contact Info

Product

Resources

About