An Analysis of Movement by Visualizing Spatio-Temporal Characteristics of Logdata

Abstract. In this study, we visualized and analyzed log data in order to analyze the spatiotemporal characteristics of “moving” and “staying activities”. As a case study, we collected and preprocessed GPS log data generated by students participating in field activities. STP (Space-Time Path) was used to visualize movement logs. “Movement” and staying places were distinguished through density-based clustering, and the time “stayed” and activities performed at staying places were examined. The problem of over-measuring time at some staying places was examined. To resolve this, the 3D Density-Based Spatial Clustering of Application with Noise (DBSCAN) was used to more accurately measure the time spent at staying places. We propose 3D DBSCAN as methodology to accurately measure spatiotemporal data. We believe this method will remain effective even as this data becomes more numerous.