[4L3-GS-4] Web intelligence

[4L3-GS-4-03] Empirical Analysis of Public Opinion on Older Drivers through SNS

〇Akito Nakanishi¹, Masao Ichikawa¹, Yukie Sano¹ (1. University of Tsukuba)

Keywords:Older driver, Twitter, Topic model, Text mining

Traffic crashes caused by older drivers have been a social issue in Japan. To prevent their traffic crashes, older drivers are encouraged to stop driving, but driving cessation might affect their physical and mental health. Therefore, there is an urgent need for evidence-based discussions about whether older people should stop driving. In this study, we extracted about 2.2 million texts on older drivers from Twitter between January 2010 and December 2021 and analyzed the polarities, potential topics, and time-series trends of the texts. We observed an increase in negative opinions on older drivers in the texts over the study period despite a decreasing trend of traffic crashes caused by older drivers. In our analysis, topics about older driver issues in the texts were divided into positive and negative ones. The positive topics appeared in the texts constantly over the study period, while the negative topics increasingly appeared, especially right after the occurrence of high-profile crashes caused by older drivers.