Over 20TB Database Identifies Hotspot Causes
The Pingtung County Government Police Bureau has built a traffic accident application platform, integrating road traffic accident investigation reports, scene diagrams, photos, video footage, violation records and other data from the past nearly 6 years, establishing a structured database of over 20TB. The platform can cross-analyze information such as accident time, location, vehicle type, collision type and cause, helping to identify accident hotspots and main causes, moving traffic governance from manual compilation and experience-based judgment toward precise data-based decision-making.
Traffic Squadron Chief 程大維 of the Pingtung County Police Bureau said that in the past accident data was huge in volume and scattered, and analysis relied on manual compilation, which was not only time-consuming but also made it difficult to quickly grasp characteristics. After the platform integrates information over the years, it can serve as a basis for improvements to road engineering, signals, signs and markings, and police deployment, and features an engineering improvement performance tracking function to continuously analyze accident data, forming a safety management cycle of analysis, improvement, tracking, and re-correction.
Added Markings at Dalian Intersection Lead to Significant Drop in Accidents
Taking the intersection of Dalian Road and Ziyou Road in Pingtung City as an example, platform analysis found that accidents mostly involved small cars and motorcycles/slow vehicles, mainly rear-end and side collisions, with causes including improper driving and failure to maintain safe distance, mostly concentrated during the morning rush hour from 8 a.m. to 9 a.m. Based on collision diagrams and the road environment, police added motorcycle waiting zones and directional lines on Dalian Road, clarifying driving and waiting positions for cars and motorcycles, reducing rear-end, sideswipe and turning conflict risks caused by weaving through traffic.
Fengrong Village Chief 陳素貞 said traffic volume at the intersection is heavy during commuting hours, with motorcycles often weaving through traffic, causing car-motorcycle conflicts. In the early stage of improvement, residents were unaccustomed due to added signals and longer waiting times, but as waiting positions and routes became clear, traffic gradually smoothed and accidents improved. Data shows overall accidents at the intersection fell from 33 in 2023 to 20 in 2026, with rear-end and side collisions down from 6 to 2.
Historical Data Supports Duty Deployment and TDX Integration
At another location, the intersection of Hesheng Road and Jiannan Road in Pingtung City, a total of 112 accidents occurred from 2018 to 2020, causing 102 injuries, averaging 3.1 accidents and 2.8 injuries per month. After analysis and improvement, from January to September 2026 there were 20 accidents and 19 injuries, averaging 2.2 accidents and 2.1 injuries per month, with the annual accident rate down 28.5%. Taking the Mid-Autumn Festival as an example, past data shows drunk-driving or nighttime accident risks are more concentrated during long holidays, which can serve as a reference for duty deployment. In the future, the platform will integrate violation and towing data, interface with the Ministry of Transportation TDX road construction events and traffic flow information, and introduce AI customer service and automatic report generation, paired with the Transportation and Tourism Department road safety analysis dashboard, to continuously improve road safety.