Abstract: Based on the network travel notes and using ROST-CM and Arcgis software, This paper analyzes the space-time characteristics of Xinjiang's self-driving tourist flow through interpretation of the semantic network, and picking up the geographic location information The following findings were obtained: (1) The temporal characteristics of Xinjiang's self- driving tourist flows are high concentration degree and uneven distribution during the year, mainly in July, August and September. So tourist season is obvious. (2) The self-driving tourist market in Xinjiang is highly concentrated in more developed regions such as Beijing, Shanghai, and Guangdong, and the eastern coastal regions and the more developed areas of the central region,including Hubei and Chongqing etc., The local self-driving tourist market in Xinjiang has a large potential, and the proportion of tourists from the neighboring provinces is not high.(3)The spatial distribution of Xinjiang's self- driving is wide,performing the attribute of patchy and zonal distribution. The spatial density is uneven, the cold and hot spots with circle distribution is obvious. (4) The spatial trajectory of self-driving tourists in Xinjiang is radial with multiple core nodes. The density of nodes in northern Xinjiang is higher than that in southern Xinjiang and eastern Xinjiang. The connecting points in northern Xinjiang are concentrated in Urumqi, Karamay, Yining etc. The points in eastern Xinjiang are Hami City, Balikun, Naomaohu Town in the Hami region. The junctions in southern Xinjiang are mainly concentrated in the Kashgar and the Hetian region, showing a trend of a large dispersion and a small concentration.

Key words: Self-driving, Tourist Flows, Temporal and Spatial Characteristics, Text Mining, Xinjiang