Analysis Of Residents’ Behavior Based On The Taxi GPS Data

Nowadays, Residents travel behavior is becoming more and more important especially for the urban traffic planning, and it has become an important reference in the city traffic construction. Since the traditional methods of residents’travel information collection usually taking large amounts of work and a long time, and the data processing is complex and boring, it turns to be no more use. The fast pace and diversification of modern urban life, making the research of residents’trip become more complex. Because of characteristics of high reliability of data, massive amounts of information, easy to get, the application of GPS is becoming more and more extensive in transportation field recently, so in this paper, we analyze the residents’travel behavior based on GPS technology. While, as a special kind of public transportation, taxi is random in travel,24hours driving; it can stop and go at anywhere and anytime, the origin and destination of taxi are completely determined by passengers, so taxi can be a good way to reflect the travel characteristics of passengers.Based on the GPS data of taxi, this paper tried to analyze the residents’travel behavior. The content of this paper mainly embodied in the following aspects:1. Putting forward the data processing standards towards the pretreatment method of GPS data and digital map matching, implementing an algorithm which based on the shortest distance through programming.2. Analyzing the characteristic of the passengers in travel time from the following aspects:the number of residents’travel, travel time and unloaded ratio.3. Discovering the hot spots of residents’travel through hierarchical clustering algorithm based on the spatial characteristics of passengers’ travel.There are some errors of GPS data because of the influence of weather, terrain and instruments, such as location errors and state display error, etc. So we need to have a pretreatment towards the GPS data and eliminating the error data, and then matching with GIS map. There are two kinds of taxi’s status,1represents loading, and0represents unloading. We could understand the behavior of passengers’get on or off the taxi through the change of taxi’s status. We could discover the characteristics of residents’travel in time and spatial distribution by calculating the time and location of passengers get on or off the taxi. For the areas where have a high frequency and density of residents’travel, we serve them as hot spots. There exist some similarities of the residents’ travel in the hot spots; we can discover the hot spots through the cluster analysis of locations of passengers’get on or get off.