Skip navigation
User training | Reference and search service

Library catalog

Content aggregators
Please use this identifier to cite or link to this item:

Title: Spatiotemporal variation of taxi demand
Authors: Rodrigues, P.
Martins, A.
Kalakou, S.
Moura, F.
Editors: Esteve Codina, Francesc Soriguera, Lídia Montero, Miquel Estrada, M. Paz Linares
Keywords: Taxi demand
Artificial Neural Network
Issue Date: 2019
Abstract: The growth of urban areas has made taxi service become increasingly more popular due to its ubiquity and flexibility when compared with, more rigid, public transportation modes. However, in big cities taxi service is still unbalanced, resulting in inefficiencies such as long waiting times and excessive vacant trips. This paper presents an exploratory taxi fleet service analysis and compares two forecast models aimed at predicting the spatiotemporal variation of short-term taxi demand. For this paper, we used a large sample with more than 1 million trips between 2014 and 2017, representing roughly 10% of Lisbon’s fleet. We analysed the spatiotemporal variation between pick-up and drop-off locations and how they are affected by weather conditions and points of interest. More, based on historic data, we built two models to predict the demand, ARIMA and Artificial Neural Network (ANN), and evaluated and compared the performance of both models. This study not only allows the direct comparison of a linear statistical model with a machine learning one, but also leads to a better comprehension of complex interactions surrounding different urban data sources using the taxi service as a probe to better understand urban mobility-on-demand and its needs.
Peer reviewed: yes
DOI: 10.1016/j.trpro.2020.03.145
ISSN: 2352-1465
Appears in Collections:BRU-CRI - Comunicações a conferências internacionais

Files in This Item:
File Description SizeFormat 
Spatiotemporal Variation of Taxi Demand.pdfVersão Editora1.09 MBAdobe PDFView/Open

FacebookTwitterDeliciousLinkedInDiggGoogle BookmarksMySpace
Formato BibTex MendeleyEndnote Currículo DeGóis 

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.