New AI-based study shows what makes roads safer

In a recent study published in the British Medical Journal (BMJ) of Injury Prevention, Dr. Quynh Nguyen, an epidemiologist and statistician at the University of Maryland School of Public Health, has utilized Google Street View (GSV) and AI tools to investigate the factors that contribute to car-related, cyclist-related, and pedestrian-related accidents. With car crashes being the leading cause of death for individuals aged between 5 and 29 years old, understanding the impact of the physical environment on these collisions is of utmost importance. This is reported by SSP.

By analyzing road features such as streetlights, greenery, sidewalks, and roadway design in locations where accidents occur, researchers were able to deduce their relationship with car accidents. Leveraging the vast amount of GSV data, precise results were obtained regarding the influence of various built elements on fatal car crashes. The study revealed that areas with higher levels of greenery, streetlights, sidewalks, and single-lane roads were associated with fewer fatal car crashes.

Sidewalks emerged as one of the most impactful features in reducing crashes, with areas boasting more sidewalks experiencing 70% fewer traffic accidents. Furthermore, regions with single-lane roads, predominantly found in rural areas, exhibited 50% fewer accidents. For pedestrians and cyclists, the presence of streetlights and stop signs offered enhanced safety and was associated with fewer car accidents involving these groups. Conversely, road construction areas had an adverse effect, resulting in increased collisions.

Xiaohe Yue, a data analyst in the UMD School of Public Health, emphasized that many public health issues can be effectively addressed, benefiting from emerging technologies and extensive data sources. The hope is that these findings will help guide transport and infrastructure policies, providing decision-makers with proven practical options to improve road safety for drivers, pedestrians, and cyclists.

Co-author Heran Mane, a data analyst working alongside Yue in the School of Public Health, expects that the study's outcomes will encourage urban planners and developers to significantly consider the built environment when designing safer streets and communities.

Dr. Nguyen envisions a new research pathway emerging from this study, emphasizing the increasing role of data science and AI in enabling larger, more efficient, and timely studies within the field of public health. There is great potential for further exploration of various built environment indicators across the United States and other countries, contributing to a deeper understanding of road safety.

As advancements in AI continue to accelerate, research efforts such as this one serve as compelling proof of its potential to enhance public health. This study stands as a testament to the positive impact technology can have on improving road safety, encouraging further exploration and innovation in this vital field.