A Taxonomy for Classifying Online Public Opinions on Transportation Services: Fid-Bau Portal

A Taxonomy for Classifying Online Public Opinions on Transportation Services

Qi, Bing / Costin, Aaron

Social media data (SMD) have the potential to become a valuable information source for transportation management. Transportation-related information exists in all forms of social media that can be applied for different aspects of transportation planning, such as evaluation of transportation services and investigating human mobility. In our preliminary research, SMD that are relevant with personal opinions on transportation services (POTS) has been extracted automatically based on a SMD collection, processing, and filtering framework for the purpose of transportation service evaluation. However, for better understanding and full exploration of the data, research on systematically classify POTS-related SMD into finer categories (e.g., transportation modes) is lacking. This paper is based on the assertation that a comprehensive taxonomy with hierarchical classification of transportation terms can better classify POTS-related SMD to evaluate different aspects of transportation services. Therefore, the focus of this paper is the development of a transportation domain taxonomy that is specific to POTS-related SMD. The taxonomy is validated by a case study in Miami-Dade County, in which Twitter is chosen as the social media. Results show the developed taxonomy have good performance in defining the semantic and syntactic information needed to identify specific information from POTS-related SMD. Significantly, the taxonomy can improve the classification of POTS-related SMD and thus enable comprehensive evaluation of the transportation services. The developed taxonomy shows the feasibility of developing other taxonomies to achieve specific objectives for SMD related service.