Evaluating the potential of online review data for augmenting traditional transportation planning performance management

Synchrony of transportation planning with revealed public preferences and expectations has traditionally been difficult to achieve. In other domains, such as manufacturing and design, frameworks have been developed using data-driven analytic processes to extract customer preferences and expectations to inform product development. This study builds from approaches to understanding customer satisfaction in other domains to investigate the potential of social media review mining for bridging the gap between understanding of neighborhood livability and conventional transportation planning decision-making practices. The novelty of the approach is in interpreting an automated user-defined translation of qualitative measures of livability by evaluating users' satisfaction of the neighborhoods through social media and enhancing the traditional approaches to defining livability planning measures (through a sentiment analysis and visualization package). The provided data-driven neighborhood satisfaction package draws the marketing structure of transportation network elements (analogous to the 'products' in other domains) with residential nodes as the center of the structure. In this study, the FastText algorithm is tested to employ a social review analysis, common neighborhood aspects are annotated, and neighborhood satisfaction is visualized and compared to the associated demographic side information. Through extracting the important aspects of the public concerns, advanced visualization helps transportation planners better understand the implications for community livability and provides insight into an approach for conducting more equitable public outreach. This approach has the potential to capitalize on residents' interests in social media outlets and to increase public engagement in the planning process by encouraging users to participate in online neighborhood satisfaction reporting. This study demonstrates a proof of concept for a social review analysis and specifies the methodology for agencies to augment planning performance management. (word count: 273).