Spatio-temporal modeling of taxi trips

Les taxis sont une composante essentielle du système de transport urbain, car ils complètent les différents modes de transport public. Contrairement aux modes transport collectif tels que le métro ou les bus, qui fonctionnent selon des itinéraires et des horaires fixes, les taxis n'ont pas d'horaires, d'itinéraires ou d'arrêts à respecter. En outre, dans les zones métropolitaines, le service de taxi est devenu un mode de transport populaire, en raison de sa rapidité et de son service porte-à-porte, de sa flexibilité et de son caractère planifiable. Une meilleure compréhension de la demande de taxis est nécessaire pour aider le secteur à améliorer l'offre de taxis afin de mieux répondre aux besoins des passagers de manière à ce que le taxi soit complémentaire aux autres modes de transport tels que le bus et le métro. En outre, la compréhension de la variation temporelle et spatiale de la demande de taxis peut aider les chauffeurs à fournir un service aux passagers au bon endroit et au bon moment. Cela peut empêcher les chauffeurs d'errer dans les rues pour trouver des passagers, ce qui entraîne une augmentation des embouteillages et de la consommation d'essence.

Abstract

Taxis are an essential component of the urban transport system as they complement the various modes of public transport. Unlike public transport modes such as metro or buses, which operate based on fixed routes and scheduled times, taxis do not adhere to any schedules, routes, or stops. Furthermore, in metropolitan areas, taxi service has become a popular mode of transportation, due to its fast and door-to-door service, flexibility, and plannability. Better understanding the taxi demand is necessary to help the industry improve the supply of taxis to better meet the needs of passengers in a way that taxi is complementary to other modes of transportation such as bus and metro. Furthermore, understanding the temporal and spatial variation of taxi demand can help drivers to provide service to passengers at the right location and time. This can prevent drivers from wandering the streets to find passengers which leads to more congestion and gasoline consumption. According to the literature, several factors can explain taxi demand, such as land use, road design, urban environment, bus stops, road length and importantly weather. Two main reasons have made forecasting taxi demand a difficult task. Firstly, many of the factors which explain the taxi demand vary both spatially and temporally, similar to the taxi demand which fluctuates over space and time. The second reason is that taxi is often used for irregular and occasional trips, which makes forecasting the pattern of its usage a difficult task.