Exploring the Built Environment Factors Influencing Town Image Using Social Media Data and Deep Learning Methods
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThe English in the manuscript needs academic proofreading.
The introduction lacks focus on the topic. For instance, the second paragraph introduces some urban-rural ratio studies, but the remainder of the study contains no information on it. The same issue is observed with mental and health topics (last sentence in this paragraph).
The literature review fails to provide a summary of the research gap on the built environment and images, making it difficult for readers to follow.
The logic and structure of the materials and methods section seem mixed. I suggest that the authors introduce the research process first and condense unnecessary methodological details. Besides, how to get the density of sidewalk?
According to the statistical description, the degree of dispersion of the town's uniqueness image appears low, with a high mean compared to the maximum value. I believe the authors should provide evidence for the representativeness of the sampled images. Considering the image source, I would question the confidence in the sampled images, especially since Little Red Book often provides unique images and recommendations for tourism.
I recommend that the authors divide the discussion and conclusion sections into two separate sections.
Comments on the Quality of English LanguageThe English in the manuscript needs academic proofreading.
Author Response
Dear reviewers,
Thank you for your patience and valuable comments. I have revised and replied one by one.
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for Authors-Introduction could be clearer and make a stronger statement about why the research matters.
-Clarify abbreviations, BE is used in the first [paragraph and not defined, it is only explained in the abstract
-I am not fully familiar with the methodology but it seems appropriate
-Conclusion could go into more detail, in particular, provide recommendations for urban planners about how this research can inform decision making
Comments on the Quality of English Language
English is generally good, and some editing would improve the text. Also consider explaining each element, sometimes terms or explanations are terse and therefore limits the audience that can understand and use the research
Author Response
Dear reviewers,
Thank you for your patience and valuable comments. I have revised and replied one by one.
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsThe study attempts an interesting piece of work, exploring the built environment factors influencing image at the town scale using social media data and deep learning methods, and arriving at some intriguing conclusions. However, there are still some problems:
l The study integrates Kevin Lynch’s theory of urban imagery, selecting relevant indicators (independent variables) from aspects such as “path”, “node”, “edge”, “district”, and “landmark”, the dependent variable is Uniqueness Image (UI), so maybe natural correlation exists, which will affect the accuracy of research conclusions.
l Kevin Lynch’s urban image primarily describes the spatial structure and behavior of the perceptual environment. Can Text Sentiment (emotions) accurately reflect tourists’ or residents’ real impressions of characteristic towns?
l Regarding the selection of research indicators, which can represent the built environment? The rationale behind the choice of indicators is unclear. Also, the selection of supplementary technical approaches and the construction of indicators, including additional references, need clarification.
l The examination of mediating effects (modulating effects) appears abrupt; moreover, is there just a simple linear relationship between positive emotions and other indicators?
l The article uses social media data but overlooks the diversity in social media use among different demographic groups.
l "5. Conclusion and Discussion" section, some content is not concise enough, and the core findings and viewpoints of the research should be summarized concisely, and improvements should be made in the conclusion and discussion..
l The study repeatedly emphasizes the mechanism of action, which is merely an elucidation of data results. The interpretative discussion of the phenomena is relatively weak, and it is suggested to enhance the exploration of causal mechanisms.
l The paper requires careful proofreading and further standardization in writing. For example, formulas and the variables involved should be italicized in the text; Figure 1 needs to be marked with approval number; “Line 310: As shown in Figure 7” is incorrectly labeled., etc.
Comments on the Quality of English LanguageThe quality of English language needs to be further strengthened
Author Response
Dear reviewers,
Thank you for your patience and valuable comments. I have revised and replied one by one.
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 3 Report
Comments and Suggestions for AuthorsThe author has carefully revised the article according to the review comments and further improvement is needed.
Comments on the Quality of English LanguageThe quality of English language still can be further improved, and it is recommended that it be polished by a professional organization.
