| Point patterns, the locations of particular events mapped as points in a study area, are one of the most common types of spatial patterns. The term 'linear points' refers to points that occur exclusively on a line or a set of lines. The location of traffic accidents in a road network is a good example of linear point data. This study develops an interactive visualization system suitable for the analysis of linear point data to identify spatial, temporal, and spatio-temporal patterns. It also demonstrates the process of developing and applying statistical procedures to confirm the findings from visual analysis.;The visualization system consists of modules which allow the user to interactively query and display the various attributes of the data. The central module in this system, the Point Visualization module, allows for easy query construction through point-and-click methods. The results of these queries are instantaneously posted to the visual display. Visualization of the temporal components of the data is facilitated by using the vertical axis of the three-dimensional display. Further exploration of the temporal dimension is made possible through the temporal slider and slide show tools.;Traffic accident data collected for Pitt County, North Carolina for the period of 1986 through 1988 were used to demonstrate the capabilities of the visualization system. The visual analysis of the data indicated a possible space-time clustering in the case of alcohol-related traffic accidents. Subsequent statistical analyses confirmed the existence of space-time clusters for some combinations of critical interpoint distances and inter-event intervals.;The visualization system developed in this study is shown to be an efficient tool for the analysis of complex spatio-temporal data. The further expansion of the system will allow the development of a more generalized system capable of processing a wider variety of data. |
