Trip Characteristics Study through Social Media Data

摘要

Sentiment analysis of social media has become a popular approach in many research areas. Most of the works were based on data collected from Twitter, which is a data source composed of text and geographic information. Many transportation researchers attempted to enrich and provide more accurate transport statistics through this approach. However, text data cannot fully provide a suitable explanatory of human's behavior. In this paper, Instagram, which is a data source containing not only text and geographic information but also photographs, is utilized for the analysis. With computer vision techniques, there is the potential to better understand human's travel behavior.