Title of Paper
Foreign tourist’s perceptions towards transport information and services using Twitter data mining: The case of typhoons in Japan
Abstract (150 Words)
Past disasters highlighted the negative perceptions of foreigner tourists with information and services. In terms of the frequency of disasters, the vulnerability of transport information and services during typhoon disasters has a great impact on foreigners, especially foreign tourists. And social media has the potential to provide an additional stream of data that used for understanding the public’s perceptions, especially during disasters. Therefore, a methodological framework for interpreting foreigners’ perceptions towards transport information and services from Twitter during typhoon disasters is presented in this research. A case study in Typhoon Faxai and Hagibis is used to implement the presented framework. A web crawler has been developed to extract tweets data posted within Japan from September 7 to 15, October 7 to 21, 2019. Second, after data pre-processing, basic text mining methods including Word Cloud and Co-Occurrence Network are conducted. Third, an embedding BERT topic model is built to generate topics of public and official side. Finally, evaluate perceived service quality quantified by sentiment scores using Topic-based VADER sentiment analysis.
Foreign tourist’s perceptions towards transport information and services using Twitter data mining: The case of typhoons in Japan
Past disasters highlighted the negative perceptions of foreigner tourists with information and services. In terms of the frequency of disasters, the vulnerability of transport information and services during typhoon disasters has a great impact on foreigners, especially foreign tourists. And social media has the potential to provide an additional stream of data that used for understanding the public’s perceptions, especially during disasters. Therefore, a methodological framework for interpreting foreigners’ perceptions towards transport information and services from Twitter during typhoon disasters is presented in this research. A case study in Typhoon Faxai and Hagibis is used to implement the presented framework. A web crawler has been developed to extract tweets data posted within Japan from September 7 to 15, October 7 to 21, 2019. Second, after data pre-processing, basic text mining methods including Word Cloud and Co-Occurrence Network are conducted. Third, an embedding BERT topic model is built to generate topics of public and official side. Finally, evaluate perceived service quality quantified by sentiment scores using Topic-based VADER sentiment analysis.