Analysis of the Human Barriers to Using Bicycles as a Means of Transportation in Developing Cities

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The topic is of great interest for sustainable mobility and cycling in particular.

In the abstract it is recommended to better emphasise the novelty of the research and to observe the template rules by excluding references in the abstract and introducing them in the various paragraphs.

It is also advisable to specify the choice of case study and possible methodological replicability

With regard to transport infrastructures and in particular cycling infrastructures, it would be useful to emphasise how, in various parts of the world, the type of cycle lane/sidewalk can guarantee greater or lesser safety and inclination to use bicycles in urban and non-urban areas.

Please consider this possibile references

1) McNeil, N., Monsere, C. M., & Dill, J. (2015). Influence of bike lane buffer types on perceived comfort and safety of bicyclists and potential bicyclists. Transportation research record, 2520(1), 132-142.

2) Campisi, T., Moslem, S., Al-Rashid, M. A., & Tesoriere, G. (2022, September). Optimal urban planning through the best–worst method: Bicycle lanes in Palermo, Sicily. In Proceedings of the Institution of Civil Engineers-Transport (Vol. 177, No. 4, pp. 203-213). Emerald Publishing Limited.

3) Guo, X., Tavakoli, A., Angulo, A., Robartes, E., Chen, T. D., & Heydarian, A. (2023). Psycho-physiological measures on a bicycle simulator in immersive virtual environments: How protected/curbside bike lanes may improve perceived safety. Transportation research part F: traffic psychology and behaviour, 92, 317-336.

In general, it is advisable to stick to the magazine's template for e.g. the layout of references and tables.

It is recommended to use similar bulleted lists and/or transform some of these into flow-charts.

How can this research improve sustainability but also accessibility and livability in urban and mobility planning?

 

 

Comments on the Quality of English Language

The English could be improved to more clearly express the research.

Author Response

Our responses to each of their comments are below. Thank you for your thoughtful and accurate review of our manuscript. These changes allowed us to improve our work, and we appreciate your help. All changes have been highlighted in red or yellow in the new version.

• Comment 1: In the abstract it is recommended to better emphasise the novelty of the research and to observe the template rules by excluding references in the abstract and introducing them in the various paragraphs. Response: The novelty of the research was described more clearly.
• Comment 2: It is also advisable to specify the choice of case study and possible methodological replicability. Response: Colombia is specified as a case study of its population.
• Comment 3: With regard to transport infrastructures and in particular cycling infrastructures, it would be useful to emphasise how, in various parts of the world, the type of cycle lane/sidewalk can guarantee greater or lesser safety and inclination to use bicycles in urban and non-urban areas. Response: Several case studies are presented that relate perceptions of safety to infrastructure. (See section 3.3)
• Comment 4: Please consider this possibile references McNeil, N., Monsere, C. M., \& Dill, J. (2015). Influence of bike lane buffer types on perceived comfort and safety of bicyclists and potential bicyclists. Transportation research record, 2520(1), 132-142; Campisi, T., Moslem, S., Al-Rashid, M. A., \& Tesoriere, G. (2022, September). Optimal urban planning through the best–worst method: Bicycle lanes in Palermo, Sicily. In Proceedings of the Institution of Civil Engineers-Transport (Vol. 177, No. 4, pp. 203-213). Emerald Publishing Limited; Guo, X., Tavakoli, A., Angulo, A., Robartes, E., Chen, T. D., \& Heydarian, A. (2023). Psycho-physiological measures on a bicycle simulator in immersive virtual environments: How protected/curbside bike lanes may improve perceived safety. Transportation research part F: traffic psychology and behaviour, 92, 317-336. Response: We thank the reviewer for recommending works that are highly relevant to our study. They have been cited in the document.
• Comment 5: In general, it is advisable to stick to the magazine's template for e.g. the layout of references and tables. Response: The document was submitted in LaTeX format to ensure its quality.
• Comment 6: It is recommended to use similar bulleted lists and/or transform some of these into flow-charts. Response:The document was submitted in LaTeX format to ensure its quality.
• Comment 7: How can this research improve sustainability but also accessibility and livability in urban and mobility planning? Response:The conclusions of the manuscript were expanded.

Reviewer 2 Report

Comments and Suggestions for Authors

This study applies logistic regression, random forest, and XGBoost to systematically assess the multiple barriers to cycling. The topic is timely for sustainable transport in developing countries, the methods are advanced. However, there are several shortcomings in study design, variable handling, and interpretation.

1. The study sample predominantly comprises young, urban, and highly educated individuals. The authors should explicitly address how potential disparities between the sample and overall population characteristics might affect the analytical validity.

2. Section 2.5 lacks theoretical justification for the classification rule of the target_bike_use variable. The manuscript should either cite established literature or present preliminary experimental evidence to support this threshold, accompanied by robustness checks.

3. ​​The conclusion asserts that “subjective barriers outweigh infrastructure,” yet Table 2 of the logistic regression shows “street lighting is poor” with OR = 1.2062 (positive), meaning individuals who perceive poor lighting are more likely to be classified as potential cyclists. In section 3.6 the authors suggest that “this indicates that people inclined to cycle are more sensitive to these environmental shortcomings.” This reverse-causality conjecture cannot be verified with cross-sectional data, and alternative explanations such as “potential cyclists avoid night riding and therefore under-report lighting problems”—cannot be ruled out. The interpretation constitutes a logical leap.​​

4. Table 1 reports a singular AUC value without clarifying whether it pertains to the training or test set. Given the implausible XGBoost AUC = 1.0, if this result is from the training set it signals severe overfitting; if it derives from a single test-set split, the estimate may still be inflated by chance. We recommend adding the mean and standard deviation of a 5-fold stratified cross-validation AUC, along with learning/validation curves, to confirm the model’s generalization performance.

Minor Comments:​​
In Figures 2 and 4, the legend overlaps with bar labels, compromising readability. Consider repositioning the legend, reducing its font size, or adopting a horizontal layout to prevent obstruction of data labels.

Table 2 contains a typographical error: “Modelo” should be corrected to “Model” in the column header.

Author Response

Our responses to each of their comments are below. Thank you for your thoughtful and accurate review of our manuscript. These changes allowed us to improve our work, and we appreciate your help. All changes have been highlighted in red or yellow in the new version.

Comment 1: The study sample predominantly comprises young, urban, and highly educated individuals. The authors should explicitly address how potential disparities between the sample and overall population characteristics might affect the analytical validity.

Response: Thank you very much for your comment. A better explanation of the demographic results has been provided in section 2.2.

Comment 2: Section 2.5 lacks theoretical justification for the classification rule of the target\_bike\_use variable. The manuscript should either cite established literature or present preliminary experimental evidence to support this threshold, accompanied by robustness checks.

Response: A more detailed explanation is presented in addition to theoretically supporting the terms of the classification rule for the target\_bike\_use variable (see Section 2.5).

Comment 3: The conclusion asserts that “subjective barriers outweigh infrastructure,” yet Table 2 of the logistic regression shows “street lighting is poor” with OR = 1.2062 (positive), meaning individuals who perceive poor lighting are more likely to be classified as potential cyclists. In section 3.6 the authors suggest that “this indicates that people inclined to cycle are more sensitive to these environmental shortcomings.” This reverse-causality conjecture cannot be verified with cross-sectional data, and alternative explanations such as “potential cyclists avoid night riding and therefore under-report lighting problems”—cannot be ruled out. The interpretation constitutes a logical leap

Response: Thank you for the helpful comment. We've updated the discussion to address this point.

Comment 4: Table 1 reports a singular AUC value without clarifying whether it pertains to the training or test set. Given the implausible XGBoost AUC = 1.0, if this result is from the training set it signals severe overfitting; if it derives from a single test-set split, the estimate may still be inflated by chance. We recommend adding the mean and standard deviation of a 5-fold stratified cross-validation AUC, along with learning/validation curves, to confirm the model’s generalization performance.

Response: We are grateful for the insightful comments provided, as they have been instrumental in refining our discussion. Accordingly, a table detailing the comparative evaluation of models performance has been added, along with the Learning Curves for each model.

____________________________________________________________________________________________________________________________________

Minor Comments:

Comment 1: In Figures 2 and 4, the legend overlaps with bar labels, compromising readability. Consider repositioning the legend, reducing its font size, or adopting a horizontal layout to prevent obstruction of data labels.

Response: Thank you very much for your comment. The figures have been modified to ensure legibility of the legends.

Comment 2 : Table 2 contains a typographical error: “Modelo” should be corrected to “Model” in the column header.

Response: We thank the reviewer for their careful reading; we have corrected the typo and thoroughly reviewed the entire document.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The theme of the work meets the aims of the Journal.

The topic of this work is interesting, mainly because it is referring to the Survey and Analysis of the barriers or factors for which, the humans avoid the use of bicycles as a mean of transportation in developing cities. Therefore, I consider this as interesting and novel Thema.

The abstract and conclusions are according with the title and clearly reflects the results of the work.

It is observed that, in the text it is used numbers as superscript which I supposed were references numbers, however the superscript number 254 do not correspond to the list of the 29 presented references at the end of all the manuscript. Therefore, it unclear which references were used. It is essential to check the references used and make the references list correctly.

Not to use superscript numbers properly, can bring confusions, that is the case in:

Page 3, line 123. The survey was distributed from January 7 to May 15, 202555. Participants.....”

Page 4, line 154 - 156. “…..numerical values from 1 to 576. "Strongly disagree" or "Totally disagree" was coded as 1; "Disagree" or "Disagree" as 2; "Neutral" or "Neutral" as 3; "Agree" as 4; and "Strongly agree" or "Totally agree" as 577. For categorical variables……”

Page 15, line 448 - 420. “…..personal convenience barriers227. This suggests that although infrastructure and road culture remain relevant, the most effective strategies to promote cycling should primarily target personal and logistical challenges228”.

Again, it is necessary to check, the references consulted for this manuscript and list them correctly.

Other important aspect to consider, is the representativity of number of surveys studied respect to the population (number of inhabitants) that the research considered, since it is studying developing cities, it would be important to support this data to give more certainty to the results achieved.

It must be reviewed the number of items studied in all the research, since it is confusing:

Page 4, lines 168 -172 . It is mentioned 34 questions:

 “The survey consisted of 34 questions related to sociodemographic information (questions 1 to 8), environmental conditions (questions 9 to 15), infrastructure and safety (questions 16 to 22), practical and logistical barriers (23-27) and personal perceptions and attitudinal barriers (28-39) and personal aspects and perceptions about bicycle use (questions 23 to 34).

 

However: In page 9, section 3.5 lines 306 – 324 “Practical and logistical barriers to bicycle use” It is listed 28 – 38 Items. And in page 10, line 360, it is mentioned the item 39: "Driving a car or being driven is more fun”.

In page 11. Section 3.6, lines 369 -390. It is necessary the review of the data reported in the table 1, since they do not correspond to those described between lines 375 to 390.

 

Author Response

Our responses to each of their comments are below. Thank you for your thoughtful and accurate review of our manuscript. These changes allowed us to improve our work, and we appreciate your help. All changes have been highlighted in red or yellow in the new version.

Comment 1: It is observed that, in the text it is used numbers as superscript which I supposed were references numbers, however the superscript number 254 do not correspond to the list of the 29 presented references at the end of all the manuscript. Therefore, it unclear which references were used. It is essential to check the references used and make the references list correctly.

Response: We want to sincerely apologize for the poor quality of the manuscript you reviewed. The issue stemmed from a required conversion from LaTeX to .doc, and I mistakenly submitted the converted file without a final proofread. I later contacted the journal to correct this, but it seems the updated file never made it to you. We have now sent the properly formatted document, addressing all of your recommendations. We are embarrassed by this oversight and apologize again for the inconvenience.

Comment 2: Not to use superscript numbers properly, can bring confusions, that is the case in:

• Page 3, line 123. The survey was distributed from January 7 to May 15, 202555. Participants.....”
• Page 4, line 154 - 156. “…..numerical values from 1 to 576. "Strongly disagree" or "Totally disagree" was coded as 1; "Disagree" or "Disagree" as 2; "Neutral" or "Neutral" as 3; "Agree" as 4; and "Strongly agree" or "Totally agree" as 577. For categorical variables……”
• Page 15, line 448 - 420. “…..personal convenience barriers227. This suggests that although infrastructure and road culture remain relevant, the most effective strategies to promote cycling should primarily target personal and logistical challenges228”.

Response: See response to comment 1

Comment 3: Again, it is necessary to check, the references consulted for this manuscript and list them correctly. 
Response: See response to comment 1

Comment 4:  Other important aspect to consider, is the representativity of number of surveys studied respect to the population (number of inhabitants) that the research considered, since it is studying developing cities, it would be important to support this data to give more certainty to the results achieved.

Response:  We appreciate your feedback. Section 2.2, "Population and Sample," has been expanded to clarify the population of interest and the sample size calculation, thereby deepening the discussion within the section.

Comment 5: Page 4, lines 168 -172 . It is mentioned 34 questions: The survey consisted of 34 questions related to sociodemographic information (questions 1 to 8), environmental conditions (questions 9 to 15), infrastructure and safety (questions 16 to 22), practical and logistical barriers (23-27) and personal perceptions and attitudinal barriers (28-39) and personal aspects and perceptions about bicycle use (questions 23 to 34), However: In page 9, section 3.5 lines 306 – 324 “Practical and logistical barriers to bicycle use” It is listed 28 – 38 Items. And in page 10, line 360, it is mentioned the item 39: "Driving a car or being driven is more fun”.

Response: Thank you for pointing this out. The text has been corrected to reflect the accurate number of 39 questions in Section 2.3 and the Results and Discussion heading.

Comment 6: In page 11. Section 3.6, lines 369 -390. It is necessary the review of the data reported in the table 1, since they do not correspond to those described between lines 375 to 390.

Response: We apologize again; this was a result of the format change. All values in Table 1 have been double-checked.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

- The research gap is not clear in the introduction section. This needs to be more emphasized.
- A good quality study of the related literature should be presented.
- In page 11, the results description does not match with the table.
- Near-perfect XGBoost metrics suggest there has been a data leakage. What precautions were taken to prevent this?
- The hyperparameters for this model is not provided.
- The explanation of Table 2 repeats the table, the derived insights can be improved.
- Please make sure to keep a consistent font.
- The class balance/imbalance needs to be mentioned.
- The conclusion needs to be improved to provide the extracted and observed insights from this study. What is the future direction?

Author Response

Our responses to each of their comments are below. Thank you for your thoughtful and accurate review of our manuscript. These changes allowed us to improve our work, and we appreciate your help. All changes have been highlighted in red or yellow in the new version.

Comment 1: The research gap is not clear in the introduction section. This needs to be more emphasized. 
Response: Following the reviewer's suggestion, we have revised the knowledge gap for clarity.

Comment 2: A good quality study of the related literature should be presented.
Response: We have introduced new studies to support our claims

Comment 3: In page 11, the results description does not match with the table.
Response: We want to sincerely apologize for the poor quality of the manuscript you reviewed. The issue stemmed from a required conversion from LaTeX to .doc, and I mistakenly submitted the converted file without a final proofread. I later contacted the journal to correct this, but it seems the updated file never made it to you. We have now sent the properly formatted document, addressing all of your recommendations. We are embarrassed by this oversight and apologize again for the inconvenience.

Comment 4: Near-perfect XGBoost metrics suggest there has been a data leakage. What precautions were taken to prevent this?
Response:  In response to your feedback, we have expanded the discussion in Section 2.5, "Data Analysis Methodology," to provide a clearer overview of the data analysis process.

Comment 6: The hyperparameters for this model is not provided.
Response: A more detailed specification of the hyperparameters is provided in Section 2.5, "Data Analysis Methodology."

Comment 7: The explanation of Table 2 repeats the table, the derived insights can be improved.
Response: The discussion has been structured to enhance clarity and offer an analysis of the findings.

Comment 8: Please make sure to keep a consistent font.
Response: Thank you for your comment. The document is formatted in LaTeX to ensure high-quality typesetting.

Comment 9: The class balance/imbalance needs to be mentioned.
Response: A more detailed specification of the class balance/imbalance is provided in Section 2.5, "Data Analysis Methodology."

Comment 10: The conclusion needs to be improved to provide the extracted and observed insights from this study. What is the future direction?
Response: The conclusions of the manuscript were expanded.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The document still contains some typos and grammatical errors. It is recommended to include more bibliographical references on lines 46-51.
Furthermore, the formatting of the bulleted lists should be checked, and a more detailed description of Figure 1 and the related graphs is needed.
In Figures 2-5, it is recommended to increase the font size and include legible legends.
Once these are corrected, the paper will be eligible for publication.

Comments on the Quality of English Language

 The English could be improved to more clearly express the research.

Author Response

Dear Prof.
Below are our responses to your comments from the second round. Thank you for your thoughtful and accurate review of our manuscript. These changes allowed us to improve our work, and we appreciate your help. All changes have been highlighted in red or yellow in the new version.

Reviewer Comments:

Comment 1: The document still contains some typos and grammatical errors. It is recommended to include more bibliographical references on lines 46-51.

Response: Two studies have been cited to support this claim. Unfortunately, studies in Colombia are virtually nonexistent

Comment 2: Furthermore, the formatting of the bulleted lists should be checked, and a more detailed description of Figure 1 and the related graphs is needed.

Response: The style of the vignettes was unified and the analysis of Figure 1 was deepened.

Comment 3: In Figures 2-5, it is recommended to increase the font size and include legible legends.

Response: The font size of the images has been increased for easier reading.

Finally, we've reviewed the entire document and corrected some typos

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

The authors have done a commendable job addressing the concerns. However, in 3.2. Environmental conditions, the bullet points start with 9. 1 to 8 cannot be found in the paper. Please make sure the numbering is correct.

Author Response

Thank you for your feedback on our manuscript. We have carefully reviewed your comments from the second round, and we greatly appreciate your thoughtful and accurate suggestions. These changes have been instrumental in improving our work. All of our revisions have been highlighted in yellow in the new version.

Reviewer Comments:

Comment 1: The authors have done a commendable job addressing the concerns. However, in 3.2. Environmental conditions, the bullet points start with 9. 1 to 8 cannot be found in the paper. Please make sure the numbering is correct.

Response: Section 3's heading clarifies that questions 1-8 of the survey collect sociodemographic information. To make the text easier to understand, the specific range of questions analyzed in each subsection is now listed in sections 3.2 through 3.5.

Author Response File: Author Response.pdf