Impact of Smartphone Activity on Pedestrian Safety: A Case Study in Seoul

As pedestrian accidents are increasing in frequency, the use of smartphones while walking has emerged as a critical issue in traffic safety. Through a case study in Seoul, this research aims to reveal the impact of smartphone activity on pedestrian safety. To this end, this study constructs a wide range of datasets, encompassing smartphone activity data, as well as transportation and social-economic data. The primary smartphone activities of pedestrian are identified through a survey of 1,000 Seoul citizens, resulting in 12 categories with 81 smartphone applications. To depict pedestrian accidents based on the collected data, this research employs deep learning models and compares their performance with conventional multiple regression models. The analysis results indicate that deep learning models utilizing smartphone data achieve better performance, particularly in estimating the rate of pedestrian accidents, which is derived without the dominant factor of the living population. Using the developed model, this research estimates pedestrian accidents under scenarios of increased smartphone usage. The estimation results suggest that accidents will likely increase with the rise of popular smartphone activities such as watching videos, riding electric scooters, or using delivery services. Rational regulations for pedestrian smartphone usage should be considered to alleviate the negative impacts of smartphone usage on pedestrian safety.