Portraying the Influence Factor of Urban Vibrancy at Street Level Using Multisource Urban Data

Round 1

Reviewer 1 Report

This paper studied the urban vibrancy and its influencing factors at the street-level. The fine-grained urban vibrancy is represented by comprehensive urban sensing data. The eight urban visual spatial features and their characteristics, greenness, openness, enclosure, imaginability, walkability, blueness, traffic flow and natural, are extracted from the street view. The six types of psychological perceptual features and their characteristics, namely wealthy, safe, lively, beautiful, boring and depression, were obtained from the scores of the volunteers. The regression model was used to express the impact of visual space and psychological perception on urban vibrancy. The research scale of this paper is finer than previous studies, and a research framework is constructed to discuss the impact of visual spatial characteristics and urban psychological perception on urban vibrancy in a quantitative manner, which has very good practical significance.

There are some problems, which should be solved before it is considered for publication.

1.     As for the selection of urban vibrancy data, it is mentioned in the paper that geotagged data such as cell phone signal data, GPS trajectory data, smart card data are effective ways to perceive urban space. However, only POI and social media data are used in the paper, so why is no geotagged data mentioned before used for experiment or verification? In addition, is it reasonable to superimpose POI and social media data to represent urban vibrancy? Is there any support from previous studies or actual data verification to prove its reasonability and validity?

2.     The Weibo check-in data used in this paper is from 2014. Other data do not specify the time. Is the data time consistent? If it's inconsistent, whether the impact of time inconsistency on the accuracy of the results has been taken into account.

3.     The extraction model of psychological perception factors is trained by the scores of volunteers, and the results are affected by the individual factors of volunteers. The number, age, gender, region and other characteristics of volunteers should be explained in the paper to prove that the selection of volunteers is reasonable as far as possible.

4.     Is the causal relationship between visual space, psychological perception and urban vibrancy reasonable? Does urban vibrancy also affect spatial and psychological characteristics?

5.     The experiments in this paper are based on the street scale. The discussion section of 5.2 further proves the superiority of the street scale over the community scale. However, how the neighborhood scale results in figure 8 are obtained needs to be supplemented in this paper.

This paper studied the urban vibrancy and its influencing factors at the street-level. The fine-grained urban vibrancy is represented by comprehensive urban sensing data. The eight urban visual spatial features and their characteristics, greenness, openness, enclosure, imaginability, walkability, blueness, traffic flow and natural, are extracted from the street view. The six types of psychological perceptual features and their characteristics, namely wealthy, safe, lively, beautiful, boring and depression, were obtained from the scores of the volunteers. The regression model was used to express the impact of visual space and psychological perception on urban vibrancy. The research scale of this paper is finer than previous studies, and a research framework is constructed to discuss the impact of visual spatial characteristics and urban psychological perception on urban vibrancy in a quantitative manner, which has very good practical significance.

There are some problems, which should be solved before it is considered for publication.

1.     As for the selection of urban vibrancy data, it is mentioned in the paper that geotagged data such as cell phone signal data, GPS trajectory data, smart card data are effective ways to perceive urban space. However, only POI and social media data are used in the paper, so why is no geotagged data mentioned before used for experiment or verification? In addition, is it reasonable to superimpose POI and social media data to represent urban vibrancy? Is there any support from previous studies or actual data verification to prove its reasonability and validity?

2.     The Weibo check-in data used in this paper is from 2014. Other data do not specify the time. Is the data time consistent? If it's inconsistent, whether the impact of time inconsistency on the accuracy of the results has been taken into account.

3.     The extraction model of psychological perception factors is trained by the scores of volunteers, and the results are affected by the individual factors of volunteers. The number, age, gender, region and other characteristics of volunteers should be explained in the paper to prove that the selection of volunteers is reasonable as far as possible.

4.     Is the causal relationship between visual space, psychological perception and urban vibrancy reasonable? Does urban vibrancy also affect spatial and psychological characteristics?

5.     The experiments in this paper are based on the street scale. The discussion section of 5.2 further proves the superiority of the street scale over the community scale. However, how the neighborhood scale results in figure 8 are obtained needs to be supplemented in this paper.

Author Response

Please see the attachment！

Author Response File: Author Response.docx

Reviewer 2 Report

It is of theoretical significance to enrich, specify, and make the concept of "urban vibrancy" more measurable. Building upon existing research results, this paper studies the correlation between street-level urban visual-spatial & psychological perception factors and urban vibrancy. Valuable and intriguing conclusions have been drawn, deserving commendation. However, as an academic article, there are some areas in which it can be improved.

1. This paper utilises multiple sources of urban data, including data from road networks, POIs, social media check-ins, street view images, etc., but a majority of these categories lacks information regarding their acquisition dates, which are recommended to be added.

2. Regarding the use of POIs and check-ins, as well as the use of the entropy weight method in constructing a comprehensive vibrancy index:

(1) Does the sample size of check-ins appear adequate when compared to that of POIs? It is recommended that the authors conduct a brief analysis.

(2) The formula for calculating entropy weights seems to be incorrect. Theoretically, the higher the entropy value, the lower the importance, whereas formula (6) gives the opposite result.

(3) It is suggested that the entropy weight calculation results of two indicators, POI and check-in, should be given explicitly so that readers can better understand the research results.

3、Regarding section 3.2.2:

(1) The model and method employed for visual-spatial factors mainly stem from the synthesis of findings in various existing literature sources. While these are appropriately acknowledged through citations, noticeable parallels exist, warranting a substantial effort to consolidate and refine them.

(2) The method employed for psychological perception factors is also based on findings in existing literature, where Figure 3 shares some similarities with the references. It is recommended to provide clarification in the caption of the figure.

(3) While the method employed for human-machine adversarial scoring comes from the findings of the existing literature and is labeled with citations, readers desire a clearer introduction of the experimental process in this specific study.

4. In table 5, it seems that MSE is not the square of RMSE?

5. Some editorial errors and suggestions:

(1) Line 248: "Rifth Ring Road" should be "Fifth Ring Road."

(2) Line 289: wrong caption for Table 1.

(3) Line 647: wrong caption for Figure 8.

(4) Both "comprehensive urban vibrancy index" and "composite urban vibrancy index" are used many times throughout the paper. It is suggested to unify them.

Finally, I would like to have a brief discussion with the authors. We believe that the concept of "urban vibrancy" includes at least three aspects: physical space (which can correspond to POIs in this study), perceptual space (which can correspond to the visual perception and emotional perception in this study), and people's activity space (which can correspond to check-ins in this study). These three types of spaces are the elements that constitute the concept of "urban vibrancy." Therefore, it seems more appropriate to describe the relationship among the elements in terms of correlation rather than causality. In this paper, terms such as "intrinsic drivers" and "determinants" are used in many places, and these terms are more inclined to describe causal relationships. Nevertheless, regression analysis can be used in both relationships. This is merely a discussion and the authors may choose to ignore it.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

This is an interesting article. The significant contribution of this article is an investigation of the impacts of psychological perspectives on urban vibrancy in addition to spatial factors. Therefore, the combination of non-spatial (psychological) and spatial factors is beautiful and makes the article an interesting use case. My overall critique is as the following:

1.       Introduction

a)       Good introduction is acknowledged

b)      For a better understanding of the readers, the term ‘urban data’ and abbreviation FCN-8s

needs to be defined/explained

c)       It is suggested to replace “perception” with "perspective" throughout the article

d)      Psychological variables are not properly defined but rather just named. E.g beauty, what is beauty referring to specifically?

e)       Line 133: The statement, “…However, field survey methods are very expensive and time-consuming, and can only capture static attributes of individuals.  Please give examples of static attributes of individuals.

f)       Line 151: Which four conditions are proposed by Jacobs to promote urban vibrancy? Please add

g)      Line 172: It is an interesting statement that location is a socio-economic factor. Any reference to support this statement?

h)      Line 185: “……for analyzing the spatiotemporal relationship between urban vibrancy and the distribution of service facilities”. What kind of service facilities? Please add

 2.       Related Work

·         Line 234: “In this paper, the improved regression-Ridge regression model is used…..” Delete the word “regression” used before  “Ridge regression model”

 3.       Data and Methods

a)       Line 253: Please add “network” after road (The foundation dataset for this study is the road network obtained from OpenStreetMap (OSM)

b)      Line 256: “The accuracy of the OSM data within the study area is relatively high”. Please add the missing reference. How would you assess the positional accuracy of open-source data? Any idea?

c)       Line 258: Why only 757 main roads were retained?  Please mention the reasons

d)      Line 273: Finally, 32,140 sampling points are selected, and 128,560 street view images are acquired for the experiment. What was the spatial and temporal resolution of the images?

e)       Line 289: The caption of Table 1 is incorrect. It should be “Type of the data sources”

o   Also correct value of Number of POIs as mentioned in Line 277 i.e. 356, 253

4.       Results

a)       Line 525: Rephrase “ Thus, we used the ridge regression model improved to minimize….” as

a.       “Thus, we used the improved ridge regression model to minimize……”

b)      Line 462: Streets with higher openness values are usually located near the boundary of the Five Rings district. What is openness values?

c)       Table 2: 14 independent variables are listed. Please explain/elaborate on all these variables

d)      Figures are not of good quality. Please change accordingly

Comments for author File: Comments.pdf

Moderate editing is required

Author Response

Please see the attachment.

Author Response File: Author Response.docx