本文以广州为案例,通过通勤高峰期间出租车GPS数据的时空挖掘,分析基于GPS起讫关联的居住地交通产生与分布特征,并讨论其所隐含的城市职住空间关系及在交通需求分析中的潜在价值。研究发现居住用地交通产生的出行距离存在ZIPF法则所表现的衰减规律,到达工业、商业金融和公共服务等用地的距离依次降低,而空间上距离呈中心城区向郊区增加的同心环模式。此外,本文还尝试从交通模型参数设定、职-住关系等方面探讨本文研究的应用方向。

One of the most important anchor point people take their activities is their home. Resident trips original-destination survey has always been the common and conventional way to acquire the data for traffic demand analysis. However, a lot of investments and surveys take a long time, manpower and money, which make it difficult to meet the command of getting traffic information timely and promptly as the city grows rapidly. Taxi GPS data is to solve these problems, since lots of researchers has applied taxi GPS data to solve the problem about transportation. This paper aims to analyze the characteristics of traffic generation and distribution of residential land in commuting peak hour, discuss the urban job-housing spatial relation it implies and explore its possible application in traffic demand analysis and traffic prediction, based on spatial-temporal data mining of taxi GPS data combining land use data of Guangzhou, China. It reveals that trip distance during commuting peak hour shows distinct decay ZIPF law. At the same time, trip distance during commuting peak hour attracted by industrial land, business financial land, and public service land decrease in order, and trip distance during commuting peak hour shows the pattern of concentric ring from central area to suburb area.