Discovering spatial and temporal patterns from taxi-based Floating Car Data: a case study from Nanjing

Shen, Jingwei; Liu, Xintao; Chen, Min

doi:10.1080/15481603.2017.1309092

Cited by 44 publications

(27 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the transportation literature, two common tessellation strategies are used. Spatial aggregations are either performed using grids, where the space is partitioned into square or rectangular grids of fixed area [3], [8], [9], or using polygons, where the space is partitioned into regular or irregular polygons of variable area [10], [11], [12]. It is common practice in the transportation literature to consider either one of these tessellation styles for spatial partitioning, without motivating the choice of the tessellation technique.…”

Section: A Related Workmentioning

confidence: 99%

Taxi Demand Forecasting: A HEDGE-Based Tessellation Strategy for Improved Accuracy

Davis

Raina

Jagannathan

2018

IEEE Trans. Intell. Transport. Syst.

View full text Add to dashboard Cite

A key problem in location-based modeling and forecasting lies in identifying suitable spatial and temporal resolutions. In particular, judicious spatial partitioning can play a significant role in enhancing the performance of location-based forecasting models. In this work, we investigate two widely used tessellation strategies for partitioning city space, in the context of real-time taxi demand forecasting. Our study compares (i) Geohash tessellation, and (ii) Voronoi tessellation, using two distinct taxi demand datasets, over multiple time scales. For the purpose of comparison, we employ classical time-series tools to model the spatio-temporal demand. Our study finds that the performance of each tessellation strategy is highly dependent on the city geography, spatial distribution of the data, and the time of the day, and that neither strategy is found to perform optimally across the forecast horizon. We propose a hybrid tessellation algorithm that picks the best tessellation strategy at each instant, based on their performance in the recent past. Our hybrid algorithm is a non-stationary variant of the wellknown HEDGE algorithm for choosing the best advice from multiple experts. We show that the hybrid tessellation strategy performs consistently better than either of the two strategies across the data sets considered, at multiple time scales, and with different performance metrics. We achieve an average accuracy of above 80% per km 2 for both datasets considered at 60 minute aggregation levels.

show abstract

Section: A Related Workmentioning

confidence: 99%

Taxi Demand Forecasting: A HEDGE-Based Tessellation Strategy for Improved Accuracy

Davis

Raina

Jagannathan

2018

IEEE Trans. Intell. Transport. Syst.

View full text Add to dashboard Cite

show abstract

“…On the other hand, it is difficult to illustrate the trajectory attributes in space and time, as they involve many aspects. It is therefore necessary to aggregate or simplify such data, and visually explore their attributes [2,31].…”

Section: Related Workmentioning

confidence: 99%

4D Time Density of Trajectories: Discovering Spatiotemporal Patterns in Movement Data

Zou

Chen

et al. 2018

IJGI

View full text Add to dashboard Cite

Abstract:Modern positioning and sensor technology enable the acquisition of movement positions and attributes on an unprecedented scale. Therefore, a large amount of trajectory data can be used to analyze various movement phenomena. In cartography, a common way to visualize and explore trajectory data is to use the 3D cube (e.g., space-time cube), where trajectories are presented as a tilted 3D polyline. As larger movement datasets become available, this type of display can easily become confusing and illegible. In addition, movement datasets are often unprecedentedly massive, high-dimensional, and complex (e.g., implicit spatial and temporal relations and interactions), making it challenging to explore and analyze the spatiotemporal movement patterns in space. In this paper, we propose 4D time density as a visualization method for identifying and analyzing spatiotemporal movement patterns in large trajectory datasets. The movement range of the objects is regarded as a 3D geographical space, into which the fourth dimension, 4D time density, is incorporated. The 4D time density is derived by modeling the movement path and velocity separately. We present a time density algorithm, and demonstrate it on the simulated trajectory and a real dataset representing the movement data of aircrafts in the Hong Kong International and the Macau International Airports. Finally, we consider wider applications and further developments of time density.

show abstract

“…Typically, among vehicle GPS data, floating car data (FCD) are likely to pave a new way for understanding individual-level travel behaviour and bring transport research to a big data era. In the last few years, FCD has been widely used in transport research, including the modelling of passenger demand [9][10][11], estimation of gas emissions [12][13][14], analysis of drivers' behaviour [15][16][17], and estimation of travel time [18][19][20]. There are several taxi FCD datasets open to the public, including Beijing datasets, New York City datasets, and so forth, and a number of studies have been undertaken around these open taxi FCD datasets [10,15].…”

Section: Introductionmentioning

confidence: 99%

Uber Movement Data: A Proxy for Average One-way Commuting Times by Car

Sun

Ren

Sun

2020

IJGI

View full text Add to dashboard Cite

Recently, Uber released datasets named Uber Movement to the public in support of urban planning and transportation planning. To prevent user privacy issues, Uber aggregates car GPS traces into small areas. After aggregating car GPS traces into small areas, Uber releases free data products that indicate the average travel times of Uber cars between two small areas. The average travel times of Uber cars in the morning peak time periods on weekdays could be used as a proxy for average one-way car-based commuting times. In this study, to demonstrate usefulness of Uber Movement data, we use Uber Movement data as a proxy for commuting time data by which commuters’ average one-way commuting time across Greater Boston can be figured out. We propose a new approach to estimate the average car-based commuting times through combining commuting times from Uber Movement data and commuting flows from travel survey data. To further demonstrate the applicability of the commuting times estimated by Uber movement data, this study further measures the spatial accessibility of jobs by car by aggregating place-to-place commuting times to census tracts. The empirical results further uncover that 1) commuters’ average one-way commuting time is around 20 min across Greater Boston; 2) more than 75% of car-based commuters are likely to have a one-way commuting time of less than 30 min; 3) less than 1% of car-based commuters are likely to have a one-way commuting time of more than 60 min; and 4) the areas suffering a lower level of spatial accessibility of jobs by car are likely to be evenly distributed across Greater Boston.

show abstract

Discovering spatial and temporal patterns from taxi-based Floating Car Data: a case study from Nanjing

Cited by 44 publications

References 45 publications

Taxi Demand Forecasting: A HEDGE-Based Tessellation Strategy for Improved Accuracy

Taxi Demand Forecasting: A HEDGE-Based Tessellation Strategy for Improved Accuracy

4D Time Density of Trajectories: Discovering Spatiotemporal Patterns in Movement Data

Uber Movement Data: A Proxy for Average One-way Commuting Times by Car

Contact Info

Product

Resources

About