This study explores the potential of an automated process in assessing the perceived quality of public spaces using Natural Language Processing (NLP) algorithms and online reviews. Over 1900 reviews were analyzed, revealing significant differences in perceptions influenced by factors such as user groups, amenities, space management, and visibility. This study found that public plazas are more welcoming and inclusive but lacking in safety and maintenance, while POPS are viewed as uninviting due to inappropriate physical design and rude personnel. In response, the study recommends policy remedies, such as stronger community relations, better design strategies, and improved information dissemination channels. While the use of the automated social media analysis method provides advantages in terms of time and labor input, improvements in accuracy require additional skill sets and data/labor input. Researchers and policymakers should still balance efficiency and accuracy while using novel methods as there is no one-size-fits-all solution to any urban issue.