旅游在线搜索与客流波动的动态关联研究——以南京钟山风景名胜区为例

Dynamic Correlation Analysis of Online Travel Information Search and Volatility in Daily Tourist Arrivals: A Case Study of Zhongshan Mountain National Park in Nanjing

波动是旅游流的重要特征和研究重点,为探索日际时间尺度下景区客流与在线信息流的波动差异和关联性,文章利用手机大数据长期监测的南京钟山景区游客量与两种客户端搜索量的3种旅游流数据,使用GARCH族方法测算不同旅游流的年内波动性,采用时差相关系数法和滚动相关系数的协动性分析和交叉相关分析方法研究了旅游流之间的整体和动态相关性,并基于VAR的脉冲响应函数研究了游前搜索对景区客流的冲击效应。分析结果表明:在日际尺度的波动上,电脑客户端(PC端)上的旅游景区搜索量序列较为平稳,移动客户端的搜索量与游客量序列具有尖峰厚尾分布和波动丛集性特征;景区客流量和移动搜索量波动均不具有长记忆性,移动搜索指数的高波动影响略长于客流量,两种波动的假日经济现象明显,均具有正向非对称性,节假日冲击促进短时集中出游,但不会带来长期客流或搜索量的显著增长;旅游搜索与客流波动在不同滞后/领先时差上具有交叉相关性,PC端搜索的波动冲击会引起游客量在滞后5期达到响应峰值,响应时序动态与移动端相反;法定节假和高温等季节性因素会引起搜索与客流相关程度在年内的动态变化。

The volatility in tourist arrivals has been the focus of tourism research. Search engines have become an important tool for destination marketing, as they can satisfy a wide range of practical needs with different marketing strategies from various platforms. The dynamic correlation analysis among information, material and passenger flow has become focus of attention in the information/digital age.As technology usage becomes more widespread, it is possible to acquire large digital footprints of tourists.This study examines the difference and correlation between tourist volume forecasting based on Baidu index and fluctuations in tourist arrivals based on mobile signal data at the interdiurnal timescale. To do this, GRACH model was applied for modeling and forecasting the volatility of yearly tourist arrivals.Pearson correlation and cross-correlation analysis were used to analyze the overall and dynamic correlation between tourist volume. The impulsive effect of online travel information search on tourist volume was investigated by using impulse response function in the VAR model. The results show the volatility of tourist volume at the destination and mobile search is similar, with"peak thick tail"distribution characteristics and volatility clusters. To a certain extent, it illustrates that tourists’ online travel information search behaviors have partially transferred to mobile device. Moreover, the volatility in tourist volume and mobile search at interdiurnal timescale has no long-term memory. However, the influence of high volatility of mobile search index is slightly longer than that of tourist volume. In addition, national holidays and seasonal factors could lead to the changes in the correlation between online search behavior and actual tourist volume. Specifically, the effects of holidays on short-term tourist flow are significant, but without significant effects on long-term tourist flow. The high temperature could constraint tourists’ potential travel intentions, and the willingness of mobile search behaviors gradually declines during winter. Lastly, the fluctuation effect of PC search causes the peak response of tourist volume at a lag of 5. The dynamic sequence of response time by PC is opposite to that of mobile devices, representing the differences of online search behavior habits corresponding to the two search indexes. Search engines have become a central part of destination marketing strategy and, as such, it is essential that different platforms with different search engines should differentiate marketing strategies.It can realize the precision marketing by applying the coupling relations between online travel information search and tourist volume at timescales, organizing a variety of tourism marketing activities, optimizing PC search index and improving online information at the early stage of holidays, and promoting the conversion rate from tourists’ online search behavior to offline tourism experience.