image: This figure illustrates how consecutive taxi orders are linked to form Taxi Mobility Chains (TMCs), revealing the movement patterns of taxis over time. Specifically: A1, A2, and A3 represent three consecutive orders from Taxi A. B1, B2, B3, and B4 represent four orders from Taxi B, where B1 and B2 are consecutive, B3 and B4 are consecutive, but B2 and B3 are not consecutive. PA1o and PA1d denote the origin and destination of order A1, respectively (with similar notation applying to other orders).
Credit: Beijing Zhongke Journal Publising Co. Ltd.
A new study published in the Journal of Geo-information Science explores the spatio-temporal patterns of taxi operations in Xiamen, China, using a novel model called the Taxi Mobility Chain (TMC). The research, led by Associate Professor Rui Xin and Master student Kuan Yang from Shandong University of Science and Technology, and Professor Jiaoe Wang (Corresponding Author) from the Chinese Academy of Sciences, analyzes the continuous movement patterns of both traditional taxis and e-hailing cars, offering a deeper understanding of their operational behaviors.
The TMC model links consecutive movements of individual taxis based on order data, overcoming the limitation of single-trip data that fails to capture comprehensive behavioral patterns and providing richer contextual information. By incorporating regional embedding techniques inspired by natural language processing, researchers were able to generate vector representations of urban areas traversed by taxis, thereby identifying regions with similar taxi mobility patterns to support subsequent analyses.
The study found that traditional taxis, which operate on a "vehicle looking for people" model, exhibit stronger spatial aggregation and higher work intensity, often shuttling back and forth in high-demand areas. In contrast, e-hailing taxis, which follow a "people looking for vehicles" model, have a broader spatial service scope, driven by individual passenger needs.
The research also highlights that the clustering results of traditional taxis are more correlated with the spatial distribution of high-density road networks, while e-hailing cars show stronger associations with administrative districts. These findings provide valuable insights for urban transport planning, helping to optimize taxi services and improve overall transportation efficiency.
For more details, please refer to the original article:
Spatio-temporal pattern analysis of taxi mobility chain supported by regional embedding.
https://www.sciengine.com/JGIS/doi/10.12082/dqxxkx.2024.240339（If you want to see the English version of the full text, please click on the “iFLYTEK Translation” in the article page.）
Article Title
Spatio-temporal Pattern Analysis of Taxi Mobility Chain Supported by Regional Embedding
Article Publication Date
25-Mar-2025