Social Media Sentiment and Service Quality Evidence in Public Transport and Passenger Mobility: A Review

Authors:
Open Access
Publisher website
Google Scholar
Abstract
Social media platforms, online reviews, app store feedback, and complaint records are important sources of user-generated textual data for understanding passenger experience in public transport and related mobility contexts. This paper reviews how social media data mining and natural language processing have been used to transform such data into evidence on service quality and satisfaction. The review focuses on public transport while selectively drawing on adjacent mobility domains, including aviation, ride-hailing, micromobility, and app-based mobility, which offer transferable methodological insights. Public transport studies are treated as the core empirical evidence base, whereas adjacent mobility- and methods-oriented studies support conceptual and methodological interpretation rather than direct claims about public transport. The review examines how textual sources are converted into analytical outputs, including sentiment, topics, aspects, emotions, and composite indicators. It shows that sentiment should not be treated as a direct measure of satisfaction but as an intermediate signal of passengers’ expressed perceptions. Key risks include sampling bias, negativity bias, linguistic ambiguity, class imbalance, weak spatial and temporal attribution, platform effects, and limited validation. The paper proposes a validation-oriented framework that maps text-derived signals to service aspects, anchors them in spatial–temporal context, triangulates them with survey, complaint, and operational evidence, and translates them into service-quality-relevant decision-support outputs.

This publication has 153 references indexed in Scilit: