BDRoadRisk: Large-Scale GPS Traffic Accident Severity Dataset

This dataset contains 10 million synthetic GPS-based traffic event records simulating road conditions on the urban road networks of Dhaka and Chittagong, Bangladesh. Each record includes 23 features covering vehicle speed, acceleration, visibility, traffic density, reaction time, weather condition, road surface, vehicle type, road quality, driver experience, alcohol influence, seatbelt usage, and GPS coordinates, and is labelled with one of three accident severity classes: No_Accident (74.41%), Minor (17.3%), and Major (8.28%). Severity labels are assigned through a context-driven probabilistic model in which road class, weather, time of day, vehicle type, road surface, and seasonal factors jointly determine per-record accident probability. Non-linear interaction terms capture amplified risk when multiple hazardous conditions co-occur. A 1.8% boundary label noise is applied at the No_Accident/Minor boundary to reflect real-world reporting ambiguity. The dataset is provided in two formats: a Hive-partitioned Apache Parquet archive compatible with Apache Spark and distributed query engines, and a single flat CSV file suitable for Pandas and R. It is intended for multi-class severity classification benchmarking, explainability research, and big data pipeline development in intelligent transportation systems.