Temporal Data Warehouse Approach for Accident Monitoring in Crossroads

Nowadays, methodologies and techniques used for road traffic data processing and analysis are continuously evolving with specific statistical analysis methods being proposed for this purpose. An example of this is crossroad monitoring, where the increasing complexity of human factors, as well as the diversity of technological aspects, affect the road control system and require suitable tools for its management. Furthermore, increased traffic is observed at crossroads where accidents constitute a major problem. Conclusively, the recording and processing of a vast amount of data relevant to accident occurrences in crossroads is of high priority, as it could assist in minimizing such occurrences by improving our understanding of and ability to confront their causes. Data warehouses are collections of data specifically designed for the support of decision making by data querying, reporting and analysis. A data warehouse supporting historical data is called a temporal data warehouse. In this paper a temporal data warehouse model is proposed and evaluated for accident monitoring in crossroads. The crossroads’ accident data warehouse is logically modeled using a temporal starnest schema which forms an integration of the star schema and the snowflake schema. The expressive power of the model is validated by a number of temporal SQL queries in Oracle 11g. Evaluation results have shown that the proposed approach is promising compared to the star schema and the snowflake schema.