Visualization Of Analyzing User’s Behavior Based On User’s Movement Data

With the development and popularity of various locating technologies, it is more easily to access user’s historical location data. The amount of user’s historical location data forms big data resources which provide a basis source of data for us to analyze people’s behaviors. Based on analyzing user’s behaviors from historical location data, it has an important realistic significance for personal services, urban construction, commercial activities, and so on. This paper focuses on the preprocessing of user’s GPS location history data, the analysis of the user’s behaviors and visualization of the user’s behaviors.In order to better analyze the user’s behavior, the works in this paper was divided into three phases. The first phase is data pre-processing. In this phase, the stay point set was found out from GPS history location points firstly. And then, individual’s important point set was acquired by using the density-based clustering methods to analyze the stay point set. After that, the semantic stay point set gotten from the Geocoding API of Baidu Map and individual’s important point set were saved the into the database. The second phase is the user’s behavior analysis. In this phase, the user’s behaviors were defined by the POI(Point Of Interest) types of the stay points, relationship among the stay points and the sequences of the stay points. And then, user’s daily behavior schemas were obtained based on finding out the stay points of daily according to POI types of the stay points and defined user’s behaviors. After that, user’s frequently non-sequence behaviors and frequently sequence behaviors were determined after analyzing all user’s historical location data by applying FP-Growth algorithm and PrefixSpan algorithm. The third phase is the design and implementation of the visual analysis tool. By using the visual analysis tool, the paper achieved the visualization of the important point set, the visualization of the daily behavior of user, the visualization of frequent non sequence behaviors and the visualization of frequent sequence behaviors separately. The figures of the visual analysis included the static pie chart, line chart, histogram, and the dynamic thermodynamic chart and trajectory chart. The results of the visual analysis were described and explained.In this paper, a visual analysis tool was designed and developed. It is a Java Web system developed by Java. The tool included four models which were the model of preprocessing trajectory data, the model of user’s behaviors, the model of the controller, and the model of the visual analysis. The model of the controller controlled and transferred data among three other models. The tool can integrated the functions of the preprocessing trajectory data, the analysis of user’s behaviors, and visualization into the same environment to facilitate using. The structure of the tool also can improve the extensible functions of the preprocessing trajectory data, the analysis of user’s behaviors, and visualization.