Tourism Forecast Based on Web Search Data and Sentiment Analysis of Social Network
Chong Zhang, Hanfei Liu, Zhiyang Chen, Tang Haiyan · The 2nd International Conference on Computing and Data Science · 2021
Rapid tourism development promotes the economy, whereas is often accompanied with resource allocation overload and serious environmental problems. One of the solutions is running tourism forecast. Most people now tend to make plans via Internet, which provides a great deal of data to forecast. This essay studies the tourist volume in Sanya, China from January 2012 to December 2019, depending on web search data and social media data. At the beginning, acquire keywords and expand them to a bigger target dataset using diction method and keyword tool. Grab their indexes afterwards. Then calculate Pearson correlation coefficient between tourist volume and search indexes, followed by picking the proper keywords. Next, implement data cleansing and sentiment analysis on texts containing keywords extracted from Sina Weibo, the dominant social media in China, to build sentiment index. Very lastly, build ARIMA model based on search index and sentiment index, to predict monthly tourist volume in the near future. Compared to traditional model not engaging both search index and sentiment analysis, our model shows higher accuracy that MAPE drops from 6.1% to 5.4% and RMSE from 14.4% to 11.1%.