A New Methodology to Study Street Accessibility: A Case Study of Avila (Spain)

Round 1

Reviewer 1 Report

This is a very interesting paper and I recommend it for publication. However, the presentation is not very clear and I think that the paper could have better reception if written in a different way. 

1) The description of complex networks in the introduction is too long and suggests that it could be a part of Methods. I have problems finding any connection between the introduction and that what is described in Methods.  The entire 3 section is mostly about the weighting nodes than working with the network. Moreover, I have little doubts if 3.3.1 is a method or a result.

2) How figure 6 relates to methods? I think that it is somehow explained in 3.1 but my concern I probably the consequence of comment 1).

3) figure 7 is completely unreadable

4) There is no Discussion section, but 4.3 seems to be a form of discussion of the results

The problem with paper is that it is not known what it is about or a relevant case study. Authors claim the method, but results and discussion concentrate on local problems. If the method is a central part of the paper the case study shall consider few scenarios and show how the method works in different conditions, what are free parameters and how changing them change the outcome etc. If the method is applied to a specific problem, than specific results should be discussed as it is in the paper.

Author Response

Thanks a lot for your review. We have revised our paper attending your recommendations as follows:

• We have rewritten the introduction, methods and experiment sections to improve the description and connection between these sections.
• We have described clearer the method and results separately.
• We have added a discussion section.
• We have solved some typos and grammar mistakes.

This is a very interesting paper and I recommend it for publication. However, the presentation is not very clear and I think that the paper could have better reception if written in a different way. 

1) The description of complex networks in the introduction is too long and suggests that it could be a part of Methods. I have problems finding any connection between the introduction and that what is described in Methods.  The entire 3 section is mostly about the weighting nodes than working with the network. Moreover, I have little doubts if 3.3.1 is a method or a result.

We absolutely agree with your recommendation.

We have rewritten the introduction and we have improved the description of the paper. Now, section 3.3 describes the method and all results have been moved to the experiments section.

2) How figure 6 relates to methods? I think that it is somehow explained in 3.1 but my concern I probably the consequence of comment 1).

We follow your suggestion (the figure belongs to the experiments section).

3) figure 7 is completely unreadable

We have improved the quality of this figure by taking the main nodes of the city (historic center). Moreover, we have changed the format to EPS for a better quality to zoom in digital format.

4) There is no Discussion section, but 4.3 seems to be a form of discussion of the results

We propose a method applied to a specific problem, and we have added a Discussion section to explain the obtained results in our case study. Moreover, we have explained the obtained results and discussed the chosen of the parameters. We have followed the specifications of the Technical Accessibility Code of Spain Buildings (alpha). We test with the maximum and minimum value of the range of these specifications, and we select the shown values in Figure 1. Finally, with respect to lambda parameters that represent the restoration levels, we have included a short explanation of how they affect.

Reviewer 2 Report

Thank you for the opportunity to review this manuscript. The authors present an approach for labelling the accessibility of urban streets and relate this to commercial data. Overall, I think this manuscript is interesting, clearly presented, and the research is competently conducted. My only suggestion in addition to carefully proofreading the manuscript again would be to extend the discussion on some of the essential ingredients and decisions the authors made in their approach. For example, for the allocation of difficulty scores, could the authors explain in a bit more detail how the parameter values (e.g. alpha for slopes) were chosen. Are all these choices arbitrary, or principled in some way (e.g. see tables 1 and 2)? In a similar vein, at least a discussion on using different values for the lambda parameters in equation 1 should be included, as well as a discussion of the mapping of difficulty to accessibility level (table 3), as the are crucial components of the research that could be quantified differently, presumably. I would also recommend the authors extend the discussion of their findings by critically reflecting on the extent to which their findings and recommendations depend on the decisions made in quantifying accessibility (as discussed above). Ideally, the authors could include a demonstration of how changing parameters, such as alpha or the lambdas, within a reasonable range would affect their findings and recommendations. This doesn’t have to be exhaustive, but some sort of demonstration for this would be useful.

Author Response

Thank you for the opportunity to review this manuscript. The authors present an approach for labelling the accessibility of urban streets and relate this to commercial data. Overall, I think this manuscript is interesting, clearly presented, and the research is competently conducted. My only suggestion in addition to carefully proofreading the manuscript again would be to extend the discussion on some of the essential ingredients and decisions the authors made in their approach. For example, for the allocation of difficulty scores, could the authors explain in a bit more detail how the parameter values (e.g. alpha for slopes) were chosen. Are all these choices arbitrary, or principled in some way (e.g. see tables 1 and 2)? In a similar vein, at least a discussion on using different values for the lambda parameters in equation 1 should be included, as well as a discussion of the mapping of difficulty to accessibility level (table 3), as the are crucial components of the research that could be quantified differently, presumably. I would also recommend the authors extend the discussion of their findings by critically reflecting on the extent to which their findings and recommendations depend on the decisions made in quantifying accessibility (as discussed above). Ideally, the authors could include a demonstration of how changing parameters, such as alpha or the lambdas, within a reasonable range would affect their findings and recommendations. This doesn’t have to be exhaustive, but some sort of demonstration for this would be useful.

Thanks a lot for your review. We have revised our paper attending your recommendations as follows:

• We have rewritten the introduction, methods and experiment sections to improve the description and connection between these sections.
• We have described clearer the method and results separately.
• We have solved some typos and grammar mistakes.
• We have added a discussion section to explain the obtained results in our case study and discuss the chosen of the parameters. We have followed the specifications of the Technical Accessibility Code of Spain Buildings (alpha). We test with the maximum and minimum value of the range of these specifications, and we select the shown values in Figure 1.
• With respect to lambda parameters that represent the restoration levels, we have included a short explanation of how they affect: these values help us to control the state of restoration of different elements of a city. If we select a low value, it means that we relax the difficulty with respect to the thresholds. In our case study, in theory, the best value is the unit because it is difficult to restore a historical area of the city of Avila. To test this hypothesis, we probe a smaller value of lambda (0.8) and we observe few changes (only 7 streets will change from red level colour to orange colour), and comparing with manual labelling, these streets could be wrong. For that, the best value in our case study is 1.

Reviewer 3 Report

I think this is an interesting approach that deserves publication
subject to addressing the minor issues listed below. In particular,
for sake of reproducibility, the authors should clarify more precisely,
where the employed data comes from and how it was collected.

Thanks to the authors for the interesting read.

Sect 2.1:
* Please also report the resolution of the height data, e.g.
  compared to SRTM

Sect. 2.1:
* l. 99: Grammar: the sentence seems incomplete
* l. 101: Grammar: "... could equally the economy... " is ungrammatical

Sect. 2.3:
* Please clarify where the data comes from.
  Ist it from OSM? Has it been collected for this study?
* What about the kerb height of sidewalks, which can be a major
  obstacle for wheelchair users?

Sect. 3.2:
* "difficulty level" might be a better word than "stage"
* Figure 1c: please use "or" instead of o with accent

Sect. 3.2.1:
* A slight reformulation of the three difficulty levels could clarify
  whether People or streets are meant to be classified.
* What would be the effect of also including the number of stairs,
  as for some people a single stair is not an obstacle, while more
  than one stairs would be?

Sect. 3.2.2:
* Table 1: Depending on the length of the slope, the difficulty is proposed
  as one of three peace-wise linear functions in the slope of the street
  with two discontinuities, e.g.,  at 10% and at 8% for ramp length <3m.
  Why not use *continuous* peace-wise linear functions?

Sect. 3.2.3:
* Sidewalks: do you consider other factors, e.g., streets declared as
  living streets do not necessarily need sidewalks?

Sect. 3.3.1:
* ll. 196-201 and Figs. 2-4: it is a bit confusing that the three
  levels i)-iii) are displayed differently in the maps. Fig. 2
  highlights accessible streets, while the other figures highlight
  inaccessible streets.

Sect. 4.:
* l. 280: "higher dense" -> "denser"

Author Response

Thanks a lot for your review. We have revised our paper attending your recommendations as follows:

• We have rewritten the introduction, methods and experiment sections to improve the description and connection between these sections.
• We have described clearer the method and results separately.
• We have added a discussion section to explain the obtained results in our case study and discuss the chosen of the parameters. We have followed the specifications of the Technical Accessibility Code of Spain Buildings (alpha). We test with the maximum and minimum value of the range of these specifications, and we select the shown values in Figure 1.
• We have solved some typos and grammar mistakes.

I think this is an interesting approach that deserves publication subject to addressing the minor issues listed below. In particular, for sake of reproducibility, the authors should clarify more precisely, where the employed data comes from and how it was collected. Thanks to the authors for the interesting read.

Sect 2.1:

* Please also report the resolution of the height data, e.g. compared to SRTM

Following your suggestion, we have compared our data with SRTM data, and the results are similar to our proposal. We have included the explanation in section 2.

Sect. 2.1:

* l. 99: Grammar: the sentence seems incomplete

* l. 101: Grammar: "... could equally the economy... " is ungrammatical

Thanks a lot for your corrections, we have solved these mistakes.

Sect. 2.3:

* Please clarify where the data comes from.

  Ist it from OSM? Has it been collected for this study?

We have rewritten the materials section to explain the data acquisition. We use OSM as our main source of data, but complement our data from Bing Maps, and finally we included some manual labelling from fieldwork activities.

* What about the kerb height of sidewalks, which can be a major obstacle for wheelchair users?

Thanks for your comments. We have included this information as follows:

“Moreover, we have labelled as problematic streets other cases such as the kerb height of sidewalks, which can be a major obstacle for wheelchair users. In case of a pedestrian street or a living street, there are not difficulty in terms of sidewalks (if the pavement is irregular, the difficulty is applied in this category)”

Sect. 3.2:

* "difficulty level" might be a better word than "stage"

* Figure 1c: please use "or" instead of o with accent

Thanks, we have solved these typos.

Sect. 3.2.1:

* A slight reformulation of the three difficulty levels could clarify  whether People or streets are meant to be classified.

* What would be the effect of also including the number of stairs, as for some people a single stair is not an obstacle, while more than one stairs would be?

Thanks, we have rewritten the section 3 (Methods and procedure) to clarify our method and the classification of the mobility restrictions, people affected, etc.

Concerning the number of stairs, we have followed the specifications of the Technical Accessibility Code of Spain Buildings (reference [38]).

Sect. 3.2.2:

* Table 1: Depending on the length of the slope, the difficulty is proposed as one of three peace-wise linear functions in the slope of the street with two discontinuities, e.g.,  at 10% and at 8% for ramp length <3m.

  Why not use *continuous* peace-wise linear functions?

Your proposal is very interesting. In this paper, we tested some other values (section 3.2.2), and finally, we have selected the specifications of the Technical Accessibility Code of Spain Buildings, basically because it is a standard in Spain.

Sect. 3.2.3:

* Sidewalks: do you consider other factors, e.g., streets declared as living streets do not necessarily need sidewalks?

We have added this information in section 3.2.2 as follows:

“Moreover, we have labelled as problematic streets other cases such as the kerb height of sidewalks, which can be a major obstacle for wheelchair users. In the case of a pedestrian street or a living street, there is no difficulty in terms of sidewalks (if the pavement is irregular, the difficulty is applied in this category).”

Sect. 3.3.1:

* ll. 196-201 and Figs. 2-4: it is a bit confusing that the three levels i)-iii) are displayed differently in the maps. Fig. 2 highlights accessible streets, while the other figures highlight inaccessible streets.

To clarify, we have rewritten the explanation in figure captions and section 4.1.1

Sect. 4.:

* l. 280: "higher dense" -> "denser"

Thanks, we have solved it.

Round 2

Reviewer 1 Report

The authors include all my comments. Now the paper is really good and I recommend it for publication.