Abstract: In order to study temporal and spatial evolution characteristics of China's road safety level and provide a basis for regional transportation planning, based on panel data of 31 provinces and autonomous regions in China from 2001 to 2018, principal component analysis method was used to calculate road safety level of each region, which was then visualized by employing ArcGIS software. Then, spatial correlation analysis was conducted on the whole and local regions by ESDA. The results show that the total number of traffic accidents in China has declined in 2001-2018, and the global agglomeration of national road traffic safety level has transformed from significant agglomeration in 2005 to the randomness in 2018. In 2005, Jilin and Sichuan were significant high-high areas, while Tibet was a significant low-high area, and Zhejiang and Guangdong were significant low-low areas. In 2010, Shandong and Henan became significant areas of high-high types, Shanxi was that of low-high types, and Guangdong became low-low area. In 2015, Sichuan, Shanxi, and Liaoning were marked areas of high-high types, but Guangdong were marked areas of low-low types. In 2018, Jiangsu was a significant high-high area, and Guangdong and Hunan were significant high-low ones.

Key words: road traffic safety level, spatial distribution, principal component analysis, panel data, exploratory spatial data analysis (ESDA)