Taxi Operation Analysis And Data Mining Based On Urban Taxi GPS Data

As an important part of urban transportation, the traffic of taxi has become an important part of the total road traffic in the city and one of the important factors that affect the distribution and forecasting accuracy of urban traffic. At present, many foreign and domestic scholars focus on the study of taxi GPS data, but mainly in the traffic state estimation, traffic behavior research, travel OD research, taxi operational management and other aspects. However, their study involved less in the analysis of time and space distribution related factors, such as road congestion level and the different types of land use, thus it can not provide relevant theory and data support for traffic management department and urban planning department.Based on taxi GPS data, this thesis mines in depth and analyses the basic operating characteristics and space-time distribution of taxis.Using data mining method to obtain the basic operational indicators of taxi, can not only provide various types of operating characteristics including empty loading rate for traffic managers, but also helps managers understand the operational status of the taxi market and provide with support for decision-making and data mining of taxi distribution afterwards. Through data mining in spatial and temporal distribution of taxi, it can perceive temporal movement on different roads and in different functional areas of taxi within the urban scale throughtout the day and also, it can analyze the relationship between the taxi distribution and road congestion degree or different functional areas, portray whether the living and working area are configured with taffic system reasonably. Also, it can provide reliable data to support managers and decision-makers for urban transportation and land use planning.This thesis preprocesses with the GPS data firstly, and uses coordinate transformation and map matching methods. On this basis, this thesis analyzes the basic indicators of the taxi operation system, including the number of operating taxies, the number of trips per day, average number of trips per day per car, average loading time per day per car, average empty loading time per day per car, empty loading rate and so on. Among them, this paper identifies the on-off-data recognition method, which is the basis of other index calculation such as the the number of trips per day and for taxi spatial and temporal distribution analysis afterwards. Then, it analyzes the temporal distribution of taxies on different roads and studies on the relationship between the congestion degree on the road and the distribution of taxi in depth.After words, it analyzes the trip characteristics of taxi in different functional areasThis thesis selects several typical areas which have clear function and gather taxies and passengers,including transportation land(Jinan Railway Station, Jinan Coach Station, Jinan West Railway Station), workplace(Qilu Software Park), residential land (Lingxiu City), commercial land(Wanda Plaza), cultural and recreational purposes (Quancheng Square), and contrast with each other using the indicators of the number of operating taxies and on-off numbers during different hours. And, for transportaion land, this paper adds another indicator of empty taxi proportion to characterize the intensity of utilization of taxi.Thirdly, it summarizes the factors affecting the level of taxi operation from the aspects of taxi operation characteristic and temporal and spatial distribution laws, and proposes improvement strategies from operational management and dispatch.At last, it designs the taxi operation management system using C# and Baidu Maps API, and displays taxi basic data, basic operational indicators and distribution of taxies visually.