Spatio-temporal mining to identify potential traff congestion based on transportation mode

Abstract

The increasing development of a city, creates density potential which could lead to traffic congestion. In recent years, the use of smartphone devices and other gadgets that have GPS (Global Positioning System) features become very commonly used in everyday activities. Previous work has built an architecture which could infer transportation mode based on GPS data. In this paper, we propose development of the previous work to detect potential traffic congestion based on transportation mode and with help from city spatial data. The data mining architecture is divided into three phases. In the first phase, we form classification model which will be used to get transportation mode information from GPS data. In the second phase, we extract spatial data, divide area into grids and divide time into several interval group. In the last phase, we use first phase result as a dataset to run in DBSCAN (Density-based spatial clustering of applications with noise) clustering algorithm for each different time interval group to know which grid area have traffic congestion potential. From this architecture, we introduced new term, cluster overlay which identify potential traffic congestion level in certain areas.