An integrated networking and visualization approach to crash data analysis is proposed and tested. A three-tier client server networking system is set up to serve as the backbone framework for data communication between the central database and distributed users. By using a local database management system, a user at the client level can view the crash data through data mapping and attribute listing and can analyze the data through nested query and sorting operations. A visualization module, based on the types, movement characteristics, and points of impact of the vehicles involved in the crash, automatically constructs the collision diagram by using data selected for visual display. Data on both intersections and roadway sections can be visualized and analyzed. The model was tested by using crash data obtained from the Ohio Department of Public Safety, and examples are presented. It was discovered that the proposed approach to crash data analysis not only is very effective but also is more accurate and time efficient than the conventional data analysis approach.