With the flourishing development of the global tourism economy, effectively utilizing user-generated content to promote the local tourism industry and capturing business opportunities at the right time have become crucial issues for industry practitioners. This study leverages data from the Tripadvisor travel website, Google Trends, and the Tourism Bureau of the Ministry of Transportation and Communications to analyze consumer reviews. By employing TF-IDF data processing, LDA topic modeling, and LightGBM technology, an accurate tourist visitor forecasting model is established. The research findings demonstrate that both Google Trends and extracted travel review variables contribute to improving the accuracy of the tourism model. Moreover, the model can achieve the purpose of predicting tourist visitor numbers using limited data in a short period of time.