Institute of Transportation, MOTC

        Understanding the movement behavior of tourists between attractions and the number of visitors at the attractions is helpful to improve the tourism policy, the public transportation services and the management department’s resource allocation. With the rapid development of technology and the increasing popularity of people using mobile phones to access the Internet, cellular data has the advantages of larger sample size, broader coverage and lower collection cost in human mobility prediction. Applying big data to explore users’ spatial-temporal trajectory can identify the potential movement patterns of users and provide subsequent applications.
        This paper developed a systematic prediction method, with conducting real-world applications in Hualien, Taiwan, and selecting 59 main attractions recommended by Tourism Bureau, by using tourists’ cellular data obtained from telecommunications company. In our study, the estimation of transition probability matrix was based on user’s trip chain and trip origin-destination matrix, then applying Markov Chain and error term method to predict number of visitors at attractions within every hour interval. The result showed that the mean absolute percentage error (MAPE) was about 20% to 30%, which indicated that the prediction method of this study was reasonable and also had a good performance of the prediction