Uncovering Bias in Objective Mapping and Subjective Perception of Urban Building Functionality: A Machine Learning Approach to Urban Spatial Perception

Round 1

Reviewer 1 Report

 

General Comments:

Throughout the paper, there are issues with vague technical descriptions and statements that need to be addressed. The authors should revise the manuscript to provide more clarity and specificity, particularly in presenting the research findings. Additionally, it is essential to appropriately reference existing methods and prior studies that form the methodological foundation. By incorporating relevant citations, the paper will be more robust and establish connections to existing research in the field more thoroughly.

 

Abstract:

The abstract provides an overview of the study, emphasizing the importance of aligning urban spaces with human needs. However, it lacks clarity and specific details about the research problem and findings. Additionally, references supporting the explanations and biases mentioned in the abstract are missing. The authors should revise the abstract to provide a concise summary of the research question, key findings, and their implications.

 

Introduction:

The introduction provides a background on objective mapping and subjective perception, but certain aspects require further clarification. Specifically, the authors should consider providing more references and citations for the chosen aspects of perception and the existence of bias. Additionally, definitions for terms such as "objective mapping" would benefit readers who may be unfamiliar with the concept. The authors should also provide more specific information about the subjective classification method based on street view images and support claims with relevant citations. For examples:

Line 34-35:”...perception of forms, structures, colors, and aesthetics...”

Why choose the four aspects, please provide clarifications, citation or reference .

 

Line 39-42: Need proper citations or references.

 

Line 60: A clear definition of ”Objective mapping” is needed.

 

Line 65-67: Please provide citation.

 

Line 81: Please further elaborate “subjective classification method based on street view images”.

 

Line 83-84: ”Current studies primarily  focus on methods stemming from a single data source.”

Why would you claim that? Need to be further justified.

 

Line 120-122: why choose “frequency density ratio and inverse distance-weighted frequency density” to classify architecture ? need proper citations or references to support this.

 

Related Works:

The literature review section is well-developed, although there are areas that can be improved. The authors should consider revising the explanation regarding the use of street view images, as they have been used for urban analysis for some time and may not be considered a new data source. Additionally, more information should be provided about Baidu for the benefit of readers who are unfamiliar with the website. The selection of features and the general statements made throughout the section should be supported by appropriate citations or references. For examples:

 

Line 154-156: the street view images(svi) has been using for urban analysis in many disciplines for quite a long time.

 

Line 160: Why select these features? it needs to be further clarified.

 

Line 169: citations.

 

Line 221 -229: The reasons for choosing “Machine Learning” should be further clarified.

 

Methodology:

The methodology section is generally well-structured, but there are areas where further clarification is necessary. The authors should provide more specific explanations and references for certain aspects, such as the choice of "low public recognition POI data" and the "WGS-84" reference. Additionally, the rationale behind the selection of distances (100 meters and 25 meters) should be justified. Definitions and explanations for terms like "K-means algorithm" should be provided to ensure readers' understanding. Moreover, the authors should include references to support decisions, such as segmenting the data into 100x100 grids.

 

Experiments and Results:

The authors should address the issue raised regarding Figure 13, which lacks clear depiction of colors due to the influence of the river. Furthermore, they should provide a reference or citation to support the claim made about contrasting with the previous figure. The authors should make necessary revisions to improve the clarity and readability of the figure.

 

 

In conclusion, this article presents an interesting study on uncovering bias in objective mapping and subjective perception of urban building functionality using a machine learning approach. However, improvements are needed in terms of clarity, specificity, and the inclusion of appropriate references to support claims and explanations. Addressing the existing review comments and revising the paper accordingly will significantly enhance its quality and contribute to the field of urban planning and design.

 

 

Author Response

 

Point 1: General Comments: Throughout the paper, there are issues with vague technical descriptions and statements that need to be addressed. The authors should revise the manuscript to provide more clarity and specificity, particularly in presenting the research findings. Additionally, it is essential to appropriately reference existing methods and prior studies that form the methodological foundation. By incorporating relevant citations, the paper will be more robust and establish connections to existing research in the field more thoroughly.

 

Response 1: Thank you for your comment. It is important to ensure that technical descriptions and statements are clear and specific to avoid confusion and improve the quality of the paper. Additionally, appropriate referencing of existing methods and prior studies is crucial to establish the methodological foundation and connect the research to existing work in the field. The specific revisions and appropriate references added have been highlighted in red in the revised manuscript, so please check the responses below.

 

Point 2: Abstract: The abstract provides an overview of the study, emphasizing the importance of aligning urban spaces with human needs. However, it lacks clarity and specific details about the research problem and findings. Additionally, references supporting the explanations and biases mentioned in the abstract are missing. The authors should revise the abstract to provide a concise summary of the research question, key findings, and their implications.

 

Response 2: Thank you for your insightful comments and suggestions on our abstract. We appreciate your feedback and agree that the abstract could benefit from more clarity and specificity. Here is our response to your comments:

1. Lack of clarity and specific details about the research problem and findings: In our revised abstract, we have explicitly stated our research problem, which is the perceptual biases and discrepancies in architectural function distribution in Shanghai's central urban district. We have also elaborated on our key findings, particularly the significant deviation between objective statistics and subjective perception of building functionalities. Please see Page 1, lines 16-20
2. Missing references supporting the explanations and biases mentioned in the abstract: We acknowledge your point about the lack of references in the abstract. However, it is common practice to avoid citations in the abstract unless absolutely necessary, as the abstract should be a standalone summary of the work. The references supporting our explanations and identified biases are provided in the main body of the paper.
3. Revision of the abstract to provide a concise summary of the research question, key findings, and their implications: We have revised our abstract to better reflect the research question, our key findings, and their implications for urban planning and sustainable urban development. Please see Page 1, lines 26-30

We hope that these revisions address your concerns effectively. We are open to further suggestions and discussions to improve our work.

 

Point 3: Introduction: The introduction provides a background on objective mapping and subjective perception, but certain aspects require further clarification. Specifically, the authors should consider providing more references and citations for the chosen aspects of perception and the existence of bias. Additionally, definitions for terms such as "objective mapping" would benefit readers who may be unfamiliar with the concept. The authors should also provide more specific information about the subjective classification method based on street view images and support claims with relevant citations. For examples:

Line 34-35:”...perception of forms, structures, colors, and aesthetics...”

Why choose the four aspects, please provide clarifications, citation or reference .

 

Response 3: The choice of forms, structures, colors, and aesthetics as aspects of perception in our study is based on established principles in environmental psychology and urban design. The form and structure of a building or an urban space significantly influence how people perceive and navigate their environment. This is a fundamental concept in urban design and architecture (Lynch, K. (1964). The Image of the City. MIT Press). Colors play a crucial role in shaping perceptions and emotions. They can influence people's feelings about a space and their behaviors within it. This is well-documented in environmental psychology literature (O'Connor, Z. (2011). Colour psychology and colour therapy: Caveat emptor. Color Research & Application, 36(3), 229-234). Aesthetics, or the visual appeal of a space, can significantly impact people's satisfaction with their environment and their desire to spend time in it. This is a key principle in urban design (Nasar, J. L. (1994). Urban design aesthetics: The evaluative qualities of building exteriors. Environment and Behavior, 26(3), 377-401). Added citations are in Page 1 line 38.

 

Point 4: Line 39-42: Need proper citations or references.

 

Response 4: Thank you for your comment. We appreciate your attention to detail and agree that lines 39-42 require appropriate citations for the claims made. We added a proper citation: Rossetti, T., Lobel, H., Rocco, V., & Hurtubia, R. (2019). Explaining subjective perceptions of public spaces as a function of the built environment: A massive data approach. Landscape and urban planning, 181, 169-178. Please see line 44.

 

Point 5: Line 60: A clear definition of ”Objective mapping” is needed.

 

Response 5: We have added a clear definition of "Objective mapping" in the revised manuscript. In the context of our study, "Objective mapping" denotes a rigorous, data-driven approach to spatially represent information. Utilizing Point of Interest (POI) data, this method ensures an empirical and quantitative depiction of architectural functions within the urban milieu. Contrasting subjective perception, which may be colored by individual biases, objective mapping ascertains an unbiased, analytical, and visual representation of urban architectural distributions. Please see page 1 lines 63-67.

 

Point 6: Line 65-67: Please provide citation.

 

Response 6: Thank you for your comment. We added a proper citation: Qiu, W., Zhang, Z., Liu, X., Li, W., Li, X., Xu, X., & Huang, X. (2022). Subjective or objective measures of street environment, which are more effective in explaining housing prices?. Landscape and Urban Planning, 221, 104358. Please see line 74.

 

Point 7: Line 81: Please further elaborate “subjective classification method based on street view images”.

 

Response 7: Thank you for your insightful comment. We agree that the term "subjective classification method based on street view images" could benefit from further clarification. In our study, the "subjective classification method based on street view images" refers to the process of categorizing urban architectural functions based on human perception derived from street view images. Unlike objective classification methods that rely on predefined categories or criteria, this method takes into account how people perceive and interpret the urban environment when they navigate through it.

In practical terms, this involves analyzing street view images and categorizing architectural functions based on visual cues. We elaborated on this method in the revised manuscript, please see line 89-91.

 

Point 8: Line 83-84: ”Current studies primarily focus on methods stemming from a single data source.” Why would you claim that? Need to be further justified.

 

Response 8: Thank you for your comment. Our claim that "current studies primarily focus on methods stemming from a single data source" is based on our review of the existing literature in the field of urban spatial perception. We observed that many studies tend to rely on one type of data source, such as surveys, satellite images, or social media data, to analyze and interpret urban spatial perception. However, we acknowledge that this is a broad statement and there are indeed studies that utilize multiple data sources. Our intention was to highlight the potential limitations of relying on a single data source and the benefits of a multi-source data approach, which is the approach we adopted in our study. In the revised manuscript, we changed the expression, please see line 93-95. “Some of the existing research in the field tends to rely on methodologies derived from a singular data source [17-18].”

 

Point 9: Line 120-122: why choose “frequency density ratio and inverse distance-weighted frequency density” to classify architecture ? need proper citations or references to support this.

 

Response 9: Thank you for your insightful comment. The choice of "frequency density ratio and inverse distance-weighted frequency density" as classification methods in our study is based on their effectiveness in capturing spatial patterns in urban environments.

1. Frequency Density Ratio: This method allows us to normalize the frequency of each architectural function in a given area, providing a relative measure that is more robust to variations in sample size across different areas. This is a common technique in spatial analysis and urban studies (Fotheringham, A. S., Brunsdon, C., & Charlton, M. (2002). Geographically Weighted Regression: The Analysis of Spatially Varying Relationships). Added citation in line .

 

2. Inverse Distance-Weighted Frequency Density: This method takes into account the spatial proximity of different architectural functions, giving more weight to functions that are closer together. This is based on the principle of spatial autocorrelation, which states that things that are closer together are more likely to be similar (Setianto, A., & Triandini, T. (2013). Comparison of kriging and inverse distance weighted (IDW) interpolation methods in lineament extraction and analysis. Journal of Applied Geology, 5(1)).

We will ensure to include these references in the revised manuscript to support our choice of these classification methods.

 

Point 10: Related Works: The literature review section is well-developed, although there are areas that can be improved. The authors should consider revising the explanation regarding the use of street view images, as they have been used for urban analysis for some time and may not be considered a new data source. Additionally, more information should be provided about Baidu for the benefit of readers who are unfamiliar with the website. The selection of features and the general statements made throughout the section should be supported by appropriate citations or references. For examples:

Line 154-156: the street view images(svi) has been using for urban analysis in many disciplines for quite a long time.

 

Response 10: Thank you for your constructive feedback on our literature review section. We appreciate your suggestions and agree that the section could benefit from further refinement. Here is our response to your comments:

1. We acknowledge your point that street view images have been used for urban analysis for some time and may not be considered a new data source. Our intention was not to present them as new but to highlight their increasing importance and utility in urban studies. We changed the expression “With the rapid development of image processing technology, street view images have emerged as a new data source for urban research” to “Given the advancements in image processing technology, the utilization of street view images in urban research has become increasingly prevalent in urban perception studies”. Please see line 156-158.
2. We understand that not all readers may be familiar with Baidu. We provided a brief description of Baidu and its relevance to our study in the revised manuscript. Baidu is a leading Chinese search engine, and its street view images provide valuable data for urban analysis in Chinese cities. Baidu Street View covers most of China's urban street-view images and is available for researchers to download. Please see line 170-171.
3. We agree that our selection of features and the general statements made throughout the section should be supported by appropriate citations or references. In the revised manuscript, all claims and choices are adequately supported by references to the relevant literature.

 

Point 11: Line 160: Why select these features? it needs to be further clarified.

 

Response 11: Thank you for your comment. In order to make our expression clearer, we modified the content of the manuscript, change " This research provides a theoretical and methodological foundation for exploring urban images, revealing the interaction between subjective perception and objective environment. A meticulous field survey was conducted, utilizing Baidu Street View (BSV) imagery of the Macau Peninsula, to assess key features of street spaces, such as openness, greenery, interface coverage, and road area ratio. The study further explored the correlation of these physical aspects with the physical and social well-being of the senior population [15]. Research was undertaken to evaluate urban walkability through the lens of cognitive mapping, deepening understanding of spatial cognition in pedestrian environments [16]." to " A series of studies are proposed to provide a theoretical and methodological basis for exploring urban imagery. A meticulous field investigation of Baidu Street View (BSV) images of the Macau Peninsula was conducted to assess key features of street space, such as openness, greenery, interface coverage, and road area ratio [26], to reveal the inter-action between subjective perception and objective environment, and to further explore the relevance of these physical aspects to the physical and social well-being of the elderly population. In addition, the perspective of cognitive maps allows the as-sessment of urban walkability and deepens the understanding of spatial perception in a walkable environment [27]." Please see lines 159-166.

 

Point 12: Line 169: citations.

 

Response 12: Thank you for your comment. We added a proper citation: J. Zhang, T. Fukuda, and N. Yabuki, “Development of a City-Scale Approach for Façade Color Measurement with Building Functional Classification Using Deep Learning and Street View Images,” ISPRS International Journal of Geo-Information, vol. 10, no. 8, Art. no. 8, Aug. 2021, doi: 10.3390/ijgi10080551. Please see line 173.

 

Point 13: Line 221 -229: The reasons for choosing “Machine Learning” should be further clarified.

 

Response 13: Thank you for your insightful comment. We chose to use machine learning in our study for several reasons, which we will clarify in our response and in the revised manuscript:

 

1. Machine learning algorithms are particularly effective at handling complex, multi-dimensional data, such as the data we collected from street view images. They can detect patterns and relationships in the data that may not be apparent or easily discernible through traditional statistical methods.
2. Machine learning models can be trained to make predictions on new, unseen data. In the context of our study, this allows us to predict architectural function distribution in other urban areas based on the patterns learned from Shanghai's central urban district.
3. Machine learning algorithms can automate the process of data analysis, making it more efficient and less prone to human error. This is particularly beneficial when dealing with large datasets, as in our study.

We also provided references to support our choice, citing studies that have successfully used machine learning in urban spatial perception research. Please see section 2.3.

 

Point 14: Methodology: The methodology section is generally well-structured, but there are areas where further clarification is necessary. The authors should provide more specific explanations and references for certain aspects, such as the choice of "low public recognition POI data" and the "WGS-84" reference. Additionally, the rationale behind the selection of distances (100 meters and 25 meters) should be justified. Definitions and explanations for terms like "K-means algorithm" should be provided to ensure readers' understanding. Moreover, the authors should include references to support decisions, such as segmenting the data into 100x100 grids.

 

Response 14: Thank you for your constructive feedback on our methodology section. We appreciate your suggestions and agree that the section could benefit from further clarification. Here is our response to your comments:

1. Low Public Recognition POI Data: We chose to include "low public recognition POI data" in our study to capture a more comprehensive view of the urban environment. These POIs, while not widely recognized, contribute to the overall character and functionality of the urban environment. We will provide more specific explanations and references for this choice in the revised manuscript.
2. WGS-84: The WGS-84 (World Geodetic System 1984) is a standard used in cartography, geodesy, and navigation. We used this reference system to ensure the accuracy and consistency of our spatial data.
3. Selection of Distances: The selection of distances (100 meters and 25 meters) was based on the typical block size in urban environments and the resolution of our street view images. an average distance of 25 meters between adjacent points, based on the urban street design method proposed by J. Gehl (J. Gehl, Cities for people. Island press, 2013). The selection of the 100-meter distance in our study is based on human visual cognition principles. Research in environmental psychology suggests that the average human field of view encompasses an area of approximately 100 meters in urban environments. This distance is often considered the limit within which people can clearly perceive and recognize details in their surroundings. Please see line 326-329
4. K-means Algorithm: The K-means algorithm is a widely used clustering algorithm that partitions data into K distinct, non-overlapping clusters. We used this algorithm to classify our data based on the selected features.
5. Segmenting the Data into 100x100 Grids: We segmented the data into 100x100 grids to facilitate the analysis and visualization of our data. This approach is commonly used in spatial analysis to transform continuous spatial data into a more manageable form. We provided references to support this decision in the revised manuscript. Please see line 451.

 

Point 15: Experiments and Results: The authors should address the issue raised regarding Figure 13, which lacks clear depiction of colors due to the influence of the river. Furthermore, they should provide a reference or citation to support the claim made about contrasting with the previous figure. The authors should make necessary revisions to improve the clarity and readability of the figure.

 

Response 15: Thank you for your valuable feedback on our experiments and results section, particularly regarding Figure 13.

We acknowledge that the depiction of colors in Figure 13 is influenced by the presence of the river, which may affect the clarity of the figure. We revised the figure to ensure that the colors are clearly distinguishable, even with the presence of the river. We consider adding a more detailed legend or using different color schemes to improve the readability of the figure. Please see revised Figure 13 in line 578.

The statement about contrasting with the previous figure was based on our observations of the data. We revised the contents and add a proper reference, and please see line 561-563.

 

Point 16: In conclusion, this article presents an interesting study on uncovering bias in objective mapping and subjective perception of urban building functionality using a machine learning approach. However, improvements are needed in terms of clarity, specificity, and the inclusion of appropriate references to support claims and explanations. Addressing the existing review comments and revising the paper accordingly will significantly enhance its quality and contribute to the field of urban planning and design.

 

Response 16: Thank you for your constructive feedback and for recognizing the potential of our study. We appreciate your suggestions for improving the clarity, specificity, and referencing in our paper. We revised the manuscript to ensure that our conclusions are clearly articulated and specific:

Our findings have significant implications for urban design and planning. They highlight the necessity of integrating both objective mapping and subjective perception to better comprehend the diverse needs and preferences of varied communities within an urban context. This integrated approach can guide the development of more inclusive, sustainable, and livable urban strategies.

Moreover, our study underscores the importance of commercial buildings' spatial dispersion and the dominant presence of residential buildings in shaping the visual landscape of urban spaces. These insights, derived from Shanghai's central urban district, can inform urban planning strategies in similar urban contexts. The implications of these discoveries lie in their potential to guide the formulation of more effective and sustainable urban strategies. Please see lines 680-698

In addition, we agree that appropriate references are crucial to support our claims and explanations. We review our manuscript to ensure that all statements are adequately supported by relevant literature.

 

We are committed to improving our manuscript based on your valuable feedback. We believe that these revisions will significantly enhance the quality of our paper and its contribution to the field of urban planning and design.

Author Response File: Author Response.pdf

Reviewer 2 Report

Initially, I would like to congratulate the authors for this excellent piece of work and urge them to proceed in future research.  The performed study deals with the detection of bias in objective and subjective perception of urban building functionality through a machine learning approach. To this end, it utilizes a three-steps methodological framework. The topic is quite interesting and well-addressed and the methods used are appropriate. Some interesting findings were also reached. The research process is clear and consistent. 

Still, some enhancements could be made to the manuscript:

1.     State-of-the-art can be better presented in Section 1. According to my opinion, it is necessary for the readers to be better introduced to the topic. Literature used in this part can also be enriched by applying researches focused on the field of urban planning/design, environmental psychology and the sense of place (i.e. check Vickas Mehta’s papers, Gibson research, Ghel & Svarre’s [2013] project about public life, Dann & Jacobson’s [2003] work about smellscape, Francesc Aletta papers about soundscape as well as Kyriakidis, et al. [2023] project about the GWA published in Cities Journal).

2.     Page 3 can be reduced. More specific, I propose to eliminate the reference in methodology, as it is going to be deeply presented in a specialized section. Concerning paper structure, I think that text should be better connected and the information can be better explained. Furthermore, although research objectives are clearly presented in the introduction, I propose to clearly mention the  research question as well as sub-queries the authors attempted to answer.

3.     Literature review section (No 2) can be enriched by referencing some more related works. 

4.     Study area is not clear. Figure 2, as well as the ones follow, have to focus on the specific area with specific limits. It is crucial to document their selection. In case they do not prefer to proceed in such demarcation, it would useful to justify the reason why this did not applied (i.e. study was about the whole city, CBD, etc). 

5.     Conclusions’ section is quite brief. According to my view, it should be enlarged (at least, one paragraph, more). Research question has to be clearly answered and the way in which the results derived by the research can be applied in other projects has to be mentioned, as well.

 

 

 

 

Author Response

 

Point 1: State-of-the-art can be better presented in Section 1. According to my opinion, it is necessary for the readers to be better introduced to the topic. Literature used in this part can also be enriched by applying researches focused on the field of urban planning/design, environmental psychology and the sense of place (i.e. check Vickas Mehta’s papers, Gibson research, Ghel & Svarre’s [2013] project about public life, Dann & Jacobson’s [2003] work about smellscape, Francesc Aletta papers about soundscape as well as Kyriakidis, et al. [2023] project about the GWA published in Cities Journal).

 

Response 1: Thank you for your valuable comments, and we cited these papers in Section 1:

1. J. Gehl, and B. Svarre, “Public space, public life: an interaction,” How to study public life, pp. 1–8, 2013.
2. F. Aletta, J. Kang, and Ö. Axelsson, “Soundscape descriptors and a conceptual framework for developing predictive soundscape models,” Landscape and Urban Planning, vol. 149, pp. 65–74, 2016.
3. C. Kyriakidis, I. Chatziioannou, F. Iliadis, A. Nikitas, and E. Bakogiannis, “Evaluating the public acceptance of sustainable mobility interventions responding to Covid-19: The case of the Great Walk of Athens and the importance of citizen engagement,” Cities, vol. 132, p. 103966, Jan. 2023, doi: 10.1016/j.cities.2022.103966.

 

Point 2: Page 3 can be reduced. More specific, I propose to eliminate the reference in methodology, as it is going to be deeply presented in a specialized section. Concerning paper structure, I think that text should be better connected and the information can be better explained. Furthermore, although research objectives are clearly presented in the introduction, I propose to clearly mention the  research question as well as sub-queries the authors attempted to answer.

 

Response 2: Thank you for your insightful feedback. We appreciate your suggestions for improving the structure and clarity of our paper.

1. Thank you for your insightful feedback. We understand your suggestion to reduce the content on page 3. However, we believe that including a brief overview of the methodology in the introduction provides a smooth transition for readers into the more detailed explanation in the subsequent section. This brief overview sets the stage for the in-depth discussion of the methodology and helps readers understand the context of our research approach.
2. We agree that the text could be better connected for a smoother flow of information. We review the entire manuscript and make necessary revisions to improve the logical flow and coherence of the text.
3. We revised the introduction to clearly articulate the main research question and the specific sub-questions we sought to answer in this study. Please see revised INTRODUCTION in page 1-3.

We appreciate your time and effort in reviewing our paper and providing these valuable suggestions.

 

Point 3: Literature review section (No 2) can be enriched by referencing some more related works.

 

Response 3: Thank you for your comment, I have added some related work which I have listed below:

24. Chen, T.; Lang, W.; Li, X. Exploring the Impact of Urban Green Space on Residents’ Health in Guangzhou, China. Journal of Urban Planning and Development 2020, 146, 05019022, doi:10.1061/(ASCE)UP.1943-5444.0000541.
25. He, N.; Li, G. Urban Neighbourhood Environment Assessment Based on Street View Image Processing: A Review of Research Trends. Environmental Challenges 2021, 4, 100090, doi:10.1016/j.envc.2021.100090.
26. Meng, L.; Wen, K.-H.; Zeng, Z.; Brewin, R.; Fan, X.; Wu, Q. The Impact of Street Space Perception Factors on Elderly Health in High-Density Cities in Macau—Analysis Based on Street View Images and Deep Learning Technology. Sustainability 2020, 12, 1799, doi:10.3390/su12051799.
27. Yu, X.; Her, Y.; Huo, W.; Chen, G.; Qi, W. Spatio-Temporal Monitoring of Urban Street-Side Vegetation Greenery Using Baidu Street View Images. Urban Forestry & Urban Greening 2022, 73, 127617.
28. Li, Y.; Yabuki, N.; Fukuda, T. Integrating GIS, Deep Learning, and Environmental Sensors for Multicriteria Evaluation of Urban Street Walkability. Landscape and Urban Planning 2023, 230, 104603, doi:10.1016/j.landurbplan.2022.104603.
29. Kamusoko, C.; Gamba, J. Simulating Urban Growth Using a Random Forest-Cellular Automata (RF-CA) Model. ISPRS International Journal of Geo-Information 2015, 4, 447–470.
30. Chaturvedi, V.; de Vries, W.T. Machine Learning Algorithms for Urban Land Use Planning: A Review. Urban Science 2021, 5, 68, doi:10.3390/urbansci5030068.

 

Point 4: Study area is not clear. Figure 2, as well as the ones follow, have to focus on the specific area with specific limits. It is crucial to document their selection. In case they do not prefer to proceed in such demarcation, it would useful to justify the reason why this did not applied (i.e. study was about the whole city, CBD, etc).

 

Response 4: Thank you for your valuable feedback on our study's area and the need for specificity in our figures. We understand the significance of clearly defining and justifying the area under study.

In response to your concerns, we revised the text to clearly delineate the study area and provide a justifiable reason for our selection. “By understanding perceptual deviations within a clearly demarcated study area (Core area within the middle ring of Shanghai), researchers can aspire to more precise and quantitative measurements of architectural functional layout and socio-economic conditions. The extraction of high-level representations from street view imagery is pivotal in this pursuit. These representations are particularly beneficial when we seek to compare objective data with subjective perceptions.” Please see lines 253-254 and 590-595.

 

Point 5: Conclusions’ section is quite brief. According to my view, it should be enlarged (at least, one paragraph, more). Research question has to be clearly answered and the way in which the results derived by the research can be applied in other projects has to be mentioned, as well.

 

Response 5: Thank you for your constructive feedback on our conclusion section. We agree that the section could benefit from further elaboration to provide a more comprehensive summary of our research and its implications.

We emphasized the impact of the discrepancy between objective mapping and subjective perception on the functional classification of urban architecture and how our machine learning approach has addressed this issue.

We also discussed how our results can be applied in other projects. Specifically, we highlight how our findings about the importance of integrating objective mapping and subjective perception can inform urban planning strategies in other urban contexts.

We expanded the conclusion to provide a more detailed summary of our research and its implications. Please see line 681-698.

 

We are committed to improving our manuscript based on your feedback.

Author Response File: Author Response.pdf

Reviewer 3 Report

No comments

Author Response

Dear reviewer,

Thank you very much

Regards

 

Reviewer 4 Report

The article starts with an interesting approach (objective and subjective analysis) but soon becomes confusing. In my humble opinion, this aproach is the most valuable point of this article and it sholudn't be lost. I strongly suggest to go on with this scientific aproach about objective/subjective analysis and its comparison.

There is a lack of precision in the objective analysis (POI, cartographic analysis, even justification of numerical allocation). It is not clearly shown why data taken from a cartography are objective and on the basis of which parameters (when it is a person who takes these data). There are also shortcomings in the subjective analysis because perception studies can only be done on the basis of surveys (and this has not been done).

Recomendation. This work can be done based on an objective analysis of the perception of the city with mobile electroencephalograms (EEG) and a subjective analysis based on surveys. But this is another article.

Author Response

 

Point 1: The article starts with an interesting approach (objective and subjective analysis) but soon becomes confusing. In my humble opinion, this approach is the most valuable point of this article and it sholudn't be lost. I strongly suggest to go on with this scientific approach about objective/subjective analysis and its comparison.

 

Response 1: Thank you for your valuable comments and insights. We wholeheartedly agree that the objective/subjective analysis approach is a fundamental aspect of this study, which we intended to highlight. We are sorry for any confusion our manuscript may have caused, and we appreciate your constructive feedback on making our approach more explicit and cohesive.

You suggested emphasizing the comparison between the objective and subjective analyses further. To address this, we revised our manuscript by adding explicit comparisons in the sections that detail our experimental results and discussions. We aim to clarify the significance of each approach and how they complement each other in understanding urban architectural functionalities. Please see lines 581-584 and lines 590-594.

 

Point 2: There is a lack of precision in the objective analysis (POI, cartographic analysis, even justification of numerical allocation). It is not clearly shown why data taken from a cartography are objective and on the basis of which parameters (when it is a person who takes these data). There are also shortcomings in the subjective analysis because perception studies can only be done on the basis of surveys (and this has not been done).

 

Response 2: Thank you for your astute observations and constructive feedback. We understand your concerns regarding the clarity and justification of our objective analysis using Points of Interest (POI) and cartographic data. We acknowledge that this aspect of the research could be further elucidated, and we appreciate your suggestion to better justify the numerical allocation methods used.

In terms of the objectivity of the cartographic data, the information collected is considered objective as it is factual and does not depend on an individual's perceptions or opinions. Nonetheless, we agree that the process of collecting these data involves human decision-making, which we will better address in the revision.

Your critique on the subjective analysis is duly noted. We agree that perception studies are typically based on surveys. For the sake of this study, we referred to 'subjective analysis' in terms of the interpretation of perceptual deviations identified through deep learning analysis of street view imagery. However, we acknowledge the importance of human-centered surveys in this context and will consider including such methodologies in our future work. Please see lines 644-648.

 

Point 3: Recomendation. This work can be done based on an objective analysis of the perception of the city with mobile electroencephalograms (EEG) and a subjective analysis based on surveys. But this is another article.

 

Response 3: Thank you for your insightful recommendation on using mobile electroencephalograms (EEG) for objective analysis and surveys for subjective analysis. We recognize the value and potential accuracy these methods could bring to our research.

However, the current paper primarily focuses on the use of machine learning methodologies, POI data, and street view imagery to understand perceptual deviations. The recommended methods would indeed be an excellent basis for another article, and we deeply appreciate your expert suggestion. We aim to integrate EEG and subjective surveys into our future research to provide a more comprehensive perspective on urban perception.

At this stage, we would prefer not to make any modifications to the current manuscript based on this feedback, as the recommended methods would involve a fundamentally different research approach. However, we are grateful for the insight and look forward to incorporating it into our future studies.

 

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

I appreciate the revisions made by the authors based on previous feedback and believe that the manuscript can be accepted in its current form pending minor formatting and spelling checks.

Pending minor formatting and spelling checks.

Reviewer 4 Report

The review is enough to be published in the present form. However I suggest to follow recommendations already done for other works in the same research line.