Complex Spatial Morphology of Urban Housing Price Based on Digital Elevation Model: A Case Study of Wuhan City, China
Round 1
Reviewer 1 Report
The work is really very interesting, but it is very confusing.
The abstract is excessively long and dispersive: it must be reduced by at least 50%.
The introduction is mixed with a bit of literature review, but all the relevant international literature together with more recent articles are missing (especially on the hedonic models). See Special Issues on Real Estate in the Sustainability journal (currently open) and on the Buildings journal (MDPI Publishing).
On the Kriging technique (at the base of GIS) the authors do not write absolutely anything.
Row 148: "To preliminarily analyze the relationship between the spatial morphology of housing price and urban spatial structure, we have designed the DEMs (see Figure 2) for housing price with four different spatial structures under multiple perspectives through simulated assignment (see Figure 1) and the interpolation model with the housing price as proposed by Meijers and Burger (2010) with four dimensions, including the extent of centralization / decentralization of integrated space and number of centers ": being said, a reader do not understand how the results of the authors in Figures 1 and 2 come out, please describe the four models mentioned and summarize the cited study.
There are no references to the segmentation of the real estate market and based on which elements the real estate data were geographically divided (row 199 and following), there is no description of the individual areas identified in figure 4.
Table 3 is quoted but is missing in the text (row 233).
RMB/m2, what is RMB? (row 234).
Row 226: "we select 318 housing projects in the city of the city that were already sold from December 2010 to December 2014 as the samples ": describe the elements of the typical projects.
Row 235: "Besides, we have calculated the GC based on the ArcGIS tool, the maximum distance and the minimum value are 7.9km, 17.6km and 1.3km respectively ": such high distances mean the sample taken as reference? Argue in a credible manner ....
The results and the conclusion of the study are unclear and very confusing: to rewrite.
Author Response
Reviewer #1
Comment 1
“The abstract is excessively long and dispersive: it must be reduced by at least 50%.”
Response:
* We rewrote the abstract based on the revision of the full text. We reduced the length of abstract and improved the expression.
Comment 2
“The introduction is mixed with a bit of literature review, but all the relevant international literature together with more recent articles are missing (especially on the hedonic models). See Special Issues on Real Estate in the Sustainability journal (currently open) and on the Buildings journal (MDPI Publishing).”
Response:
* We added the latest literature (MDPI Publishing, eg: Bilbao‐Terol, et al., 2017; Sun, et al., 2017; Won & Lee, 2017, Rosato, et al., 2017) on the application of the Hedonic model in the literature review section (see Row 42‐51). Further, we added additional explanations of the basic theory of the Hedonic pricing model (Xiao, 2017) in 2.1 (see Row 113‐115).
* We removed some of the references for DEM applications in the natural sciences such as Xiong, et al. (2014) ，Callow, et al. (2007), Claessens, et al. (2005), Schwind, et al. (2009).
Comment 3
“On the Kriging technique (at the base of GIS) the authors do not write absolutely anything.”
Response:
* We added literature review of Kriging and Cokriging (eg: Anselin, et al., 2005; Cellmer, 2014; Chica Olmo, 2007; ChicaOlmo, et al., 2013; Dubin, 1998; Martínez, et al., 2000; Montero‐Lorenzo, et al., 2009; José M. Montero, et al., 2015; José María Montero, et al., 2017; Olmo, 1995). Further, we pointed out the advantages, disadvantages, and uses of Kriging method (see Row 57‐69).
* We indicated that: “Previous research on urban housing prices mainly uses Kriging and Co‐Kriging techniques for price prediction, while the application of geovisual analytics is rare, especially for the analysis and visualization of nonlinear complex urban spatial structure and spatial morphology” (see Row 103‐106). Also, we emphasized that: “Prior studies that have indicated both the importance and limitations of spatial interpretation methods used for housing price prediction. In this paper, we did not use spatial interpolation methods for housing price prediction, but employed the Kriging interpolation method for producing the DEM of urban housing price.”(see Row 368‐371).
* We added the explanation of the principles and theoretical basis of the Ordinary Kriging method (Krige, 1953; Matheron, 1963; Cressie, 1993) in this paper. We just chose Ordinary Kriging as the interpretation tool for producing DEM. How to analyze the complex spatial morphology of urban housing price through a novel method based on DEM is the research focus of this paper. Therefore, we have not presented a detailed introduction to the calculation process of the Ordinary Kriging method.
Comment 4
As to Row 148 in original manuscript: “Reader do not understand how the results of the authors in Figures 1 and 2 come out, please describe the four models mentioned and summarize the cited study.”
Response:
* We adjusted the structure of this manuscript. We removed the sections related with Figures 1 and 2 in the original manuscript. We plan to rewrite another paper specifically for this topic in the future. Removing this part also helps the reader to focus on the ʺDEM‐based geovisual analytics of urban housing priceʺ we proposed in this study.
Comment 5
“There are no references to the segmentation of the real estate market and based on which elements the real estate data were geographically divided (row 199 and following), there is no description of the individual areas identified in figure 4.”
Response:
* We have added the introductions to the study area (see Row 172‐174; Row 188‐194). We introduced the geographical division of the three sectors, Hankou, Hanyang, and Wuhang, in Wuhan City, and figured out the functional differences between three regions. Further, we indicated the spatial heterogeneity of the three sectors in urban housing price and urban residential land price by referring to prior literature (Hu, et al., 2012; Hu, et al., 2016; Zhang, et al., 2015; Jiao and Liu, 2010; Han, 2004 )
Comment 6
“Table 3 is quoted but is missing in the text (row 233).”
Response:
* We solved the missing problem of Table 3.
Comment 7
“RMB/m2, what is RMB? (row 234).”
Response:
* We stated that “we used the unit selling price (in CNY ¥ ▪ m‐2) of estates project to reflect urban housing price” (see Row 220‐221).
* PS: The renminbi (Ab.: RMB; sign: ¥; code: CNY) is the official currency of the Peopleʹs Republic of China.
Comment 8
As to Row 226 in original manuscript: “describe the elements of the typical projects.”
Response:
* We rewrote section 3.2 to provide a supplementary explanation of the selection and processing of samples (see Row 214‐225; Row 231‐239).
* In addition, we emphasized that this study focuses on a new DEM‐based geovisual analytics method for housing prices (see revised Introduction and Conclusion). The producing of DEM in this study is only to provide a basis for the application example of this method. And the housing price reflected by DEM is not a real market transaction price, and there is rare significance of price prediction.
Comment 9
As to Row 235 in original manuscript: “such high distances mean the sample taken as reference? Argue in a credible manner ....”
Response:
* Please see our response to the Comment 1 of Reviewer 2.
Comment 10
“The results and the conclusion of the study are unclear and very confusing: to rewrite.”
Response:
* We rewrote the Concluding Discussion (see Row 365‐416).
Reviewer 2 Report
This study employs an interesting method to analyze the spatial pattern urban housing prices. My comments are as follow:
Is the sample size of 318 big enough to create a surface for the entire city? Are you able to validate the accuracy of the estimates based on actual newer housing prices?
How the three analytical methods (water level, section cutting, and belt floating) are formed?
What is the theoretical basis of the method design?
What is the practical use of this method? Does it help inform any decision? Why it is advantageous compared to more traditional methods? This should be discussed in the final section.
Author Response
Reviewer #2
Comment 1
“Is the sample size of 318 big enough to create a surface for the entire city? Are you able to validate the accuracy of the estimates based on actual newer housing prices?”
Response:
* The application of Kriging method should be difficult to reflect the individual differences in the housing market. In order to avoid these shortcomings, we have classified and screened the samples. We rewrote section 3.2 to further explain the choice of samples (see Row 214‐225; Row 231‐239).
* Finally, the 318 remaining samples (positions), including 313230 housing units (cover an area of about 32.26 million square meters) were used for geovisual analysis (see Row 247‐248). Further, compared with the prior research, Hu, et al. (2016) collected 460 samples successfully developed a spatial interpolation map for the urban residential land price by using the Bayesian Kriging method in the same study area (Wuhan).
* Kriging method has the advantages of achieving quicker and better global prediction when an amount of the sample data is limited (Montero‐Lorenzo, et al., 2009) (see Row 68‐69), which is one of the reasons this method has been widely applied.
Comment 2
“How the three analytical methods (water level, section cutting, and belt floating) are
formed?”
Response:
* We rewrote the section 2.2 (see Row 152‐167). We presented basic formulas, morphologies, and specific descriptions of the DEM‐based analysis method for urban housing price, and introduced how the implementation of these methods.
Comment 3
“What is the theoretical basis of the method design?”
Response:
* We explained that the idea of the methods we designed have referred to the application of DEM in other disciplines such as hydrology and geosciences, and we further discussed its advantages and limitations in the conclusions of the study (see revised Introduction, Methods, and Conclusion).
* We added the references about the basis of our designed method such as Gichamo, et al. (2012), Schäppi, et al. (2010), Tate, et al. (2002).
Comment 4
“What is the practical use of this method? Does it help inform any decision? Why it is advantageous compared to more traditional methods? This should be discussed in the final section.”
Response:
* We rewrote the Concluding Discussion (see Row 365‐416).
Round 2
Reviewer 1 Report
The manuscript is a re-submission of a previous paper, already examined from this reviewer.
Respect my previous comments, at this stage il paper is undoubtedly very improved. The authors have modified the paper according to my previous comments, almost totally. My only suggestions are now related about a more exaustive description (in analitical terms) about section 2.1 and 2.2. For the rest, the paper is fine and it is almost publishable.
Author Response
Reviewer #1
Comment 1
“Respect my previous comments, at this stage il paper is undoubtedly very improved. The authors have modified the paper according to my previous comments, almost totally. My only suggestions are now related about a more exaustive description (in analitical terms) about section 2.1 and 2.2. For the rest, the paper is fine and it is almost publishable.”
Response:
* We have added the details to explain the OK calculation method in section 2.1(see Row 193-206).
* We have added some details to explain the implementation process of Water-flooding, Section-cutting and Belt-floating methods in 2.2 (see Row 229-240).
***End Response***
Author Response File: 
Author Response.docx
Reviewer 3 Report
This research takes Wuhan City as a case to use the Kriging interpolation method to analyze the spatial morphologic characteristics of the digital elevation model (DEM) of urban housing prices and geovisualize the results from the transport and waterfront landscape perspectives. This research is interesting readable since it applies a new approach, 3D analyst module in GIS environment to visualize and detect the spatial variance of urban housing surface in relationships with transport and waterfront landscape factors. It has great potentials to show its significant merits to be published. Yet, it is particularly important that this paper could carefully analyze the geographical context within which the research was carried out (dwelling characteristics, transport infrastructure characteristics, natural resources and neighborhood amenities etc.). Such an analysis will enable this paper to better explain the complex spatial morphology of urban housing price, especially complex features of multiple peaks and fluctuations. In addition, a great discussion of the policy implications is expected to add into the paper. Specifically, I have the following specific observations regarding this paper.
First, the introduction seems appropriate, with a clear justification of application of three-dimensional analysis in investigating the complex spatial morphology of urban housing price. Yes, it is lacking an identification of insights on urban housing prices and relative urban development issues provided by three dimensional analysis over two-dimensional analysis.
Second, the literature review part is missing. The literature reviews on housing prices and monocentric and polycentric urban structures are needed. Also, the literature reviews on hedonic housing price model related to waterfront resources and transport amenities are also needed. Thus, it may require updating some recent references on the above subjects in recent years, especially related to Wuhan, such as Huang and Yin (2015), Xu et al. (2016), and Xu and Zhang (2016).
Third, the sections of models, methods, and study area are well written, but in Table 2, housing price measurement unit need be indicated since it is obvious that the unit is not 1 as default.
Fourth, the results are well presented. Yet, some results on complex features of multiple peaks and fluctuations are not sufficiently discussed. The sufficient discussions are needed to understand how these peaks and fluctuations are related to the geographical context within which the research was carried out (polycentric urban structure, dwelling characteristics, transport infrastructure characteristics, natural resources and neighborhood amenities etc.)
Fifth, the conclusions are a good review of the article. Yet, it is lacking a discussion of the policy implications of the results on guiding land allocation for urban planning and development as is mentioned in the abstract.
Last, this paper can be benefited from another round of editing, such as typos. In line 92, should it be “important” perspective instead of “impotent” perspective?
Author Response
Reviewer #3
Comment 1
“First, the introduction seems appropriate, with a clear justification of application of three-dimensional analysis in investigating the complex spatial morphology of urban housing price. Yes, it is lacking an identification of insights on urban housing prices and relative urban development issues provided by three dimensional analysis over two-dimensional analysis.”
Response:
* As suggested, we have given insights on urban housing prices and relative urban development issues provided by three dimensional analysis over two-dimensional analysis (see Row 106-108).
Comment 2
“Second, the literature review part is missing. The literature reviews on housing prices and monocentric and polycentric urban structures are needed. Also, the literature reviews on hedonic housing price model related to waterfront resources and transport amenities are also needed. Thus, it may require updating some recent references on the above subjects in recent years, especially related to Wuhan, such as Huang and Yin (2015), Xu et al. (2016), and Xu and Zhang (2016).”
Response:
* We have cited the literature on monocentric and polycentric urban structures, especially using housing price as evidence of polycentric urban structures, such as Burger and Meijers (2012), Vasanen (2012), Wen and Tao (2015), Qin and Han (2013), Mori (2016).
* We have added references to the Hedonic housing price model application for waterfront resources and transport amenities, especially to Hedonic model applications in Wuhan in recent years, such as Huang and Yin (2015), Xu et al. (2016), and Xu and Zhang (2016).
Comment 3
“Third, the sections of models, methods, and study area are well written, but in Table 2, housing price measurement unit need be indicated since it is obvious that the unit is not 1 as default.”
Response:
* Thanks to the reviewer’s careful work, we modified the units in Table 2 and corrected the data.
Comment 4
“Fourth, the results are well presented. Yet, some results on complex features of multiple peaks and fluctuations are not sufficiently discussed. The sufficient discussions are needed to understand how these peaks and fluctuations are related to the geographical context within which the research was carried out (polycentric urban structure, dwelling characteristics, transport infrastructure characteristics, natural resources and neighborhood amenities etc.)”
Response:
* As suggested, we added a section to compare and discuss the results of empirical research with the previous research on Wuhan City (see Row 691-697).
* As we explained in the first round of revisions, our paper focuses on the development of a novel approach for geovisual analytics of urban housing based on DEM. The empirical analysis of the study area is preferred to provide an examples of the application of the method. Therefore, we have not specifically tested the impact of various spatial and geographic elements on housing prices. Correspondingly, we added an explanation of the limitations of this article in the conclusion of the article (see Row 688-691).
Comment 5
“Fifth, the conclusions are a good review of the article. Yet, it is lacking a discussion of the policy implications of the results on guiding land allocation for urban planning and development as is mentioned in the abstract.”
Response:
* We have added a discussion of the results and methods of empirical analysis for the potential role of local government in land, housing management and urban planning (see Row 697-700).
* We have revised the corresponding section in the Abstract.
* As mentioned earlier, since the focus of this paper is not on policy analysis of the study case, and this paper also lacks the foundation and support for in-depth policy analysis, we have not made the further discussion on policy implications. However, we will focus on policy analysis in our next research and paper writing plan.
Comment 6
“Last, this paper can be benefited from another round of editing, such as typos. In line 92, should it be “important” perspective instead of “impotent” perspective?”
Response:
* Thanks to the reviewer’s meticulous work, we have corrected the spelling of "important" and asked a native English speaking colleague to check the language expression and spelling of the full text.
***End Response***
Author Response File: Author Response.docx
Round 3
Reviewer 3 Report
The revision addressed issues mentioned in the review comments. I understand that the objective of the empirical analysis is not to specifically test the impact of various spatial and geographic elements on housing prices, but I still think that a brief discussion on mechanisms behind peaks and fluctuations will be helpful to its merits.
