Exploring the Citywide Human Mobility Patterns of Taxi Trips through a Topic-Modeling Analysis,Journal of Advanced Transportation

当前位置： X-MOL 学术 › J. Adv. Transp. › 论文详情

Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)

Exploring the Citywide Human Mobility Patterns of Taxi Trips through a Topic-Modeling Analysis
Journal of Advanced Transportation ( IF 2.0 ) Pub Date : 2021-07-20 , DOI: 10.1155/2021/6697827
Hui Xiong ₁ , Kaiqiang Xie ₁ , Lu Ma ₂ , Feng Yuan ₂ , Rui Shen ₂

Affiliation

Understanding human mobility patterns is of great importance for a wide range of applications from social networks to transportation planning. Toward this end, the spatial-temporal information of a large-scale dataset of taxi trips was collected via GPS, from March 10 to 23, 2014, in Beijing. The data contain trips generated by a great portion of taxi vehicles citywide. We revealed that the geographic displacement of those trips follows the power law distribution and the corresponding travel time follows a mixture of the exponential and power law distribution. To identify human mobility patterns, a topic model with the latent Dirichlet allocation (LDA) algorithm was proposed to infer the sixty-five key topics. By measuring the variation of trip displacement over time, we find that the travel distance in the morning rush hour is much shorter than that in the other time. As for daily patterns, it shows that taxi mobility presents weekly regularity both on weekdays and on weekends. Among different days in the same week, mobility patterns on Tuesday and Wednesday are quite similar. By quantifying the trip distance along time, we find that Topic 44 exhibits dominant patterns, which means distance less than 10 km is predominant no matter what time in a day. The findings could be references for travelers to arrange trips and policymakers to formulate sound traffic management policies.

更新日期：2021-07-20

点击分享查看原文

点击收藏

公开下载

阅读更多本刊最新论文本刊介绍/投稿指南11