The impact of rainfall on the temporal and spatial distribution of taxi passengers

Chen, Dandan; Zhang, Yong; Gao, Liangpeng; Geng, Nan; Li, Xuefeng

doi:10.1371/journal.pone.0183574

Cited by 12 publications

(9 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This could be explained by the fact that discretionary passengers are more affected by rain than 635 commuters (Changnon, 1996), i.e., people are obliged to go out for their daily duties on weekdays and therefore might adapt their travel modes, while on weekends they can opt to stay at home under heavy rain. In addition, similarly to what Chen et al (2017) found when studying the impact of rainfall on taxi use, we also found a trend of more vehicular use under rainy weather, and less walking in heavier rain.…”

supporting

confidence: 87%

Weather effects on human mobility: a study using multi-channel sequence analysis

Brum-Bastos

Long

Demšar

2018

Computers, Environment and Urban Systems

View full text Add to dashboard Cite

Widespread availability of geospatial data on movement and context presents opportunities for 10 applying new methods to investigate the interactions between humans and weather conditions. Understanding the influence of weather on human behaviour is of interest for diverse applications, such as urban planning and traffic engineering. The effect of weather on movement behaviour can be explored through Context-Aware Movement Analysis (CAMA), which integrates movement geometry with its context. More specifically, we use multi-channel sequence analysis (MCSA) to represent a 15 person's movement as a multi-dimensional sequence of states, describing either the type of movement or the state of the environment throughout time. Similar movement patterns can then be identified by comparing and aligning mobility sequences. In this paper we apply CAMA and MCSA to explore weather effects on human movement patterns. Data from a GPS tracking study in a Scottish town of Dunfermline are linked to weather data and converted into multi-channel sequences which are 20 clustered into groups of similar behaviours under specific weather typologies. Our findings show that the CAMA + MCSA method can successfully identify the response of commuters to variations in environmental conditions. We also discuss our findings on how travel modes and time spent at different places are affected by meteorological conditions, mainly wind, but also rainfall, daylight duration, temperature, comfort and relative humidity. 25

show abstract

supporting

confidence: 87%

Weather effects on human mobility: a study using multi-channel sequence analysis

Brum-Bastos

Long

Demšar

2018

Computers, Environment and Urban Systems

View full text Add to dashboard Cite

show abstract

“…Other exogenous factors, such as inclement weather can affect short-term demand differently by mode, as noted in the case of road traffic (Thakuriah and Tilahun 2013), public transport (Tang and Thakuriah 2012), and bicycling (Meng et al 2016). The significant relationship of inclement weather to taxi use (Chen et al 2017) and ride-hailing (Brodeur and Nield 2017) has also been pointed out. Other long-term socio-economic factors, for example, income levels or employment activity, can also affect travel demand.…”

Section: The Ride-hailing Short-term Demand Forecasting Problemmentioning

confidence: 94%

Short-Term Prediction of Demand for Ride-Hailing Services: A Deep Learning Approach

Chen¹,

Thakuriah

Ampountolas

2021

J. Big Data Anal. Transp.

View full text Add to dashboard Cite

As ride-hailing services become increasingly popular, being able to accurately predict demand for such services can help operators efficiently allocate drivers to customers, and reduce idle time, improve traffic congestion, and enhance the passenger experience. This paper proposes UberNet, a deep learning convolutional neural network for short-time prediction of demand for ride-hailing services. Exploiting traditional time series approaches for this problem is challenging due to strong surges and declines in pickups, as well as spatial concentrations of demand. This leads to pickup patterns that are unevenly distributed over time and space. UberNet employs a multivariate framework that utilises a number of temporal and spatial features that have been found in the literature to explain demand for ride-hailing services. Specifically, the proposed model includes two sub-networks that aim to encode the source series of various features and decode the predicting series, respectively. To assess the performance and effectiveness of UberNet, we use 9 months of Uber pickup data in 2014 and 28 spatial and temporal features from New York City. We use a number of features suggested by the transport operations and travel behaviour research areas as being relevant to passenger demand prediction, e.g., weather, temporal factors, socioeconomic and demographics characteristics, as well as travel-to-work, built environment and social factors such as crime level, within a multivariate framework, that leads to operational and policy insights for multiple communities: the ride-hailing operator, passengers, third-part location-based service providers and revenue opportunities to drivers, and transport operators such as road traffic authorities, and public transport agencies. By comparing the performance of UberNet with several other approaches, we show that the prediction quality of the model is highly competitive. Further, Ubernet’s prediction performance is better when using economic, social and built environment features. This suggests that Ubernet is more naturally suited to including complex motivators of travel behavior in making real-time demand predictions for ride-hailing services.

show abstract

“…These travel times for ride-sharing programs are also affected by weather. For example, Chen et al [2017] showed that taxi demand decreased with rainfall during non-rush hours but increased with rainfall during rush hours. As a result, both the probability of a real request and the times to serve will be influenced by weather conditions.…”

Section: Practical Implicationsmentioning

confidence: 99%

The Perils of Learning Before Optimizing

Cameron¹,

Hartford²,

Lundy³

et al. 2021

Preprint

View full text Add to dashboard Cite

Formulating real-world optimization problems often begins with making predictions from historical data (e.g., an optimizer that aims to recommend fast routes relies upon travel-time predictions). Typically, learning the prediction model used to generate the optimization problem and solving that problem are performed in two separate stages. Recent work has showed how such prediction models can be learned end-to-end by differentiating through the optimization task. Such methods often yield empirical improvements, which are typically attributed to end-to-end making better error tradeoffs than the standard loss function used in a two-stage solution. We refine this explanation and more precisely characterize when end-to-end can improve performance. When prediction targets are stochastic, a two-stage solution must make an a priori choice about which statistics of the target distribution to model-we consider expectations over prediction targets-while an end-to-end solution can make this choice adaptively. We show that the performance gap between a two-stage and end-to-end approach is closely related to the price of correlation concept in stochastic optimization and show the implications of some existing POC results for our predict-then-optimize problem. We then consider a novel and particularly practical setting, where coefficients in the objective function depend on multiple prediction targets. We give explicit constructions where (1) two-stage performs unboundedly worse than end-to-end; and (2) two-stage is optimal. We identify a large set of real-world applications whose objective functions rely on multiple prediction targets but which nevertheless deploy two-stage solutions. We also use simulations to experimentally quantify performance gaps.Recently there has been a lot of interest in the "end-to-end" approach of training a predictive model in a way that minimizes loss on the downstream optimization task [

show abstract

The impact of rainfall on the temporal and spatial distribution of taxi passengers

Cited by 12 publications

References 29 publications

Weather effects on human mobility: a study using multi-channel sequence analysis

Weather effects on human mobility: a study using multi-channel sequence analysis

Short-Term Prediction of Demand for Ride-Hailing Services: A Deep Learning Approach

The Perils of Learning Before Optimizing

Contact Info

Product

Resources

About