Exploring the Influences of Safety and Energy Expenditure Parameters on Cycling

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This is a well written paper and provides interesting findings for the bike shares in many cities, specifically exploring factors on how affected the bike shares are. I have some comments and suggestions for the authors as follows:

1. In general, travel choice modeling is to figure out the proportion among travel modes (i.e., auto, public transportation, bike….etc.). The study simply seems to estimate the bike share itself as a dependent variable with other predictors, which became a general regression modeling. I would suggest deleting the term “Travel choice model” in the title and throughout the text.

2. In abstract, please contain the key finding from the Table 3 and concise the purpose of study.

3. Line 81, what does the ATP stand for? Please use full text first, then abbreviations come later. I hope to check this throughout the text.

4. Eq(1), Eq(3); the term ‘m’ and ‘M’ could be used as one style for the readers.

5. In eq. 4: I don't see how the EE is estimated. Is it equivalent as the W of eq.1? Please make sure on those in the text.

6. Line 315: Why did you fix only 2 variables throughout the entire city in the BS model. I would prefer to use a different set of variables by city category as shown in Table 2.

7. Line 383~386: It seemed the CLS parameter more dominant than EE, since the magnitude of coefficient is much larger.

8. Line 258: Typo error Table 2 -> 1.

Author Response

Reviewer 1

This is a well written paper and provides interesting findings for the bike shares in many cities, specifically exploring factors on how affected the bike shares are. I have some comments and suggestions for the authors as follows:

R101. In general, travel choice modeling is to figure out the proportion among travel modes (i.e., auto, public transportation, bike….etc.). The study simply seems to estimate the bike share itself as a dependent variable with other predictors, which became a general regression modeling. I would suggest deleting the term “Travel choice model” in the title and throughout the text.

R. Thank you for your interesting observation. The term “Travel choice model” have been deleted throughout the text and substituted with “Regression Model”

 

R102. In abstract, please contain the key finding from the Table 3 and concise the purpose of study.

R. Thank you for your observation. The key finding of table 3 are added and also the purpose of the study in a more concise way.

 

R103. Line 81, what does the ATP stand for? Please use full text first, then abbreviations come later. I hope to check this throughout the text.

R. Thank you for your observation. The abbreviation is declared in the whole text.

 

R104. Eq(1), Eq(3); the term ‘m’ and ‘M’ could be used as one style for the readers.

R. Thank you for your observation. The term related to the mass of cyclist, bike and accessories is now coherent in the whole paper.

 

R105. In eq. 4: I don't see how the EE is estimated. Is it equivalent as the W of eq.1? Please make sure on those in the text.

R. Thank you for your observation. The Eq. 4 is added in order to show how Energy expenditure is obtained.

 

R106. Line 315: Why did you fix only 2 variables throughout the entire city in the BS model. I would prefer to use a different set of variables by city category as shown in Table 2.

R. Thank you for your interesting observation The authors agree. The other variables (as selected in Table 2) are added to the model. The original idea was only to highlight Cycle Lane Supply and Energy Expenditure parameters.

 

R107. Line 383~386: It seemed the CLS parameter more dominant than EE, since the magnitude of coefficient is much larger.

R. Thank you for your observation. The models are a little bit changed after the first-round review.

 

R108. Line 258: Typo error Table 2 -> 1.

R. Thank you for your observation. The error has been corrected.

Reviewer 2 Report

Comments and Suggestions for Authors

The study presents a significant contribution to bicycle infrastructure planning, highlighting the importance of safety and energy expenditure in the decision to cycle. The external validation of the results and the exploration of other variables in future research can further strengthen the study's conclusions. Therefore, I recommend the paper towards acceptance. 

Constructive criticisms and suggestions for improvements:

More literature reviews should be added to support the research.

I don't think the content in the background sector is proper.  The authors may add some backgrounds or data of Italian transportation development.

Exploring other variables, such as socioeconomic and psychological factors, can enrich the research and offer a more complete view of modal choice.

The results section could be more focused on the most relevant findings and their practical implications for urban planners and transportation engineers.

The text can be improved in terms of clarity and conciseness, especially in sections with complex or redundant language.

The use of graphs, tables, and other data visualizations can be further explored to facilitate understanding of the results and relationships between variables.

The discussion section could be more in-depth, exploring the implications of the results for future research and for the practice of urban planning. Example:

Discuss how the results can be used to identify critical areas for cyclist safety and to promote cycling as a sustainable mode of transportation.

Author Response

Reviewer 2

The study presents a significant contribution to bicycle infrastructure planning, highlighting the importance of safety and energy expenditure in the decision to cycle. The external validation of the results and the exploration of other variables in future research can further strengthen the study's conclusions. Therefore, I recommend the paper towards acceptance. 

Constructive criticisms and suggestions for improvements:

R201. More literature reviews should be added to support the research.

R. Thank you for your constructive criticisms and suggestions for improvements: the authors have appreciated these ones. More literature review is added in the paper.

 

R202. I don't think the content in the background sector is proper.  The authors may add some backgrounds or data of Italian transportation development.

R. Thank you for your observation. Some data on the current state of the Italian transportation development are added.

 

R203. Exploring other variables, such as socioeconomic and psychological factors, can enrich the research and offer a more complete view of modal choice.

R. Thank you for your interesting observation. The authors agree. The other variables (as selected in Table 2) are added to the model.

 

R204. The results section could be more focused on the most relevant findings and their practical implications for urban planners and transportation engineers.

R. Thank you for your observation. Relevant findings and practical implications are added.

 

R205. The text can be improved in terms of clarity and conciseness, especially in sections with complex or redundant language.

R. Thank you for your observation. The text has been improved in term of clarity and conciseness.

 

R206. The use of graphs, tables, and other data visualizations can be further explored to facilitate understanding of the results and relationships between variables.

R. Thank you for your observation. More graphs are added in order to explain the relationships between variables.

 

R207. The discussion section could be more in-depth, exploring the implications of the results for future research and for the practice of urban planning. Example: Discuss how the results can be used to identify critical areas for cyclist safety and to promote cycling as a sustainable mode of transportation.

R. Thank you for your interesting observations. Some other considerations are added in order to explore implications of the results for future research and for the application to urban planning.

Reviewer 3 Report

Comments and Suggestions for Authors

This paper is a well written paper. It is easily readable and understandable. The paper explores the influences of safety and energy expenditure on cycling Travel Choice models. The applied methodologies are straightforward and sound reasonable. The datasets seem to be sufficient. The findings may be useful for the further research works and for the practice.

However, some gramma errors and typos must be checked out before a publication.

Some comments in details:

Eq. (4): Please clarify how the parameter EE is calculated exactly. Is EE = MET = VO2 in Eq. (3)?

Line 80: The abbreviation ATP is not declared in advance.

Line 301: The abbreviation MET is not declared in advance.

Author Response

Reviewer 3

This paper is a well written paper. It is easily readable and understandable. The paper explores the influences of safety and energy expenditure on cycling Travel Choice models. The applied methodologies are straightforward and sound reasonable. The datasets seem to be sufficient. The findings may be useful for the further research works and for the practice.

However, some gramma errors and typos must be checked out before a publication.

Some comments in details:

 

R301. Eq. (4): Please clarify how the parameter EE is calculated exactly. Is EE = MET = VO2 in Eq. (3)?

R. Thank you for your observation. The Eq. 4 is added in order to show how Energy expenditure is obtained. The number related to equation is a little bit changed.

 

R302. Line 80: The abbreviation ATP is not declared in advance.

R. Thank you for your observation. The abbreviation is declared in the introduction.

 

R303. Line 301: The abbreviation MET is not declared in advance.

R. Thank you for your observation. The abbreviation of MET is declared in the text.

Reviewer 4 Report

Comments and Suggestions for Authors

I found the abstract vague. I do not think that they estimate a travel choice model but a regression model that tries to explain bike usage share. I see no clear findings in the abstract.

What I find strange in this paper is that they state that indeed many determinants explain cycling, but they choose only two in their regression: cycle lane supply and energy expenditure. Why? I would control for the other parameters mentioned in Table 1 besides mean slope and cycle lane. And I would add parameters such as ‘active cycling policies within the city, the share of student population in the city… things like that. I find their reason not include these parameters very odd: is due to the fact that these variables seem quite biased (why biased???? Seems perfectly normal and measurable determinant to me) and very dependent on the Italian context, characterized by geographical differences in income and car insurance distribution ( in the North, where the bike share is higher than the South, the mean income is higher and the car insurance is lower than the other regional areas (yes indeed so I would control for these determinants).

I also do not quite understand why they do not just take mean slope into account in their regression. Why is the regression so ‘difficult’ with EE when the mean slope is the only difference between the cities?

They make three models per size of the city.  They continue with the model 50,000-100,000. Why only with this model? And is this choice ‘validation’ as they call it?

The conclusion is a bit vague. What do they mean with ‘it’ in line 443. I would invite them to discuss who can use their model. What kinds of policy insights can come out of their model?  

 

 

Comments on the Quality of English Language

no comments

Author Response

Reviewer 4

R401. I found the abstract vague. I do not think that they estimate a travel choice model but a regression model that tries to explain bike usage share. I see no clear findings in the abstract.

R.Thank you for your observation. The abstract has been modified. In the abstract it is highlighted that a regression model is evaluated. Also, the findings are highlighted.  The abs

 

R402. What I find strange in this paper is that they state that indeed many determinants explain cycling, but they choose only two in their regression: cycle lane supply and energy expenditure. Why? I would control for the other parameters mentioned in Table 1 besides mean slope and cycle lane. And I would add parameters such as ‘active cycling policies within the city, the share of student population in the city… things like that. I find their reason not include these parameters very odd: is due to the fact that these variables seem quite biased (why biased???? Seems perfectly normal and measurable determinant to me) and very dependent on the Italian context, characterized by geographical differences in income and car insurance distribution ( in the North, where the bike share is higher than the South, the mean income is higher and the car insurance is lower than the other regional areas (yes indeed so I would control for these determinants).

R. Thank you for your interesting observation. The authors agree. The other variables (as selected in Table 2) are added to the model.

 

R403. I also do not quite understand why they do not just take mean slope into account in their regression. Why is the regression so ‘difficult’ with EE when the mean slope is the only difference between the cities?

R. Thank you for your interesting observation. EE is more complex than terrain related parameters because it is associated to terrain related characteristics, human capabilities (related to gender and age) and type of mode used (c-bike or e-bike).

 

R404. They make three models per size of the city.  They continue with the model 50,000-100,000. Why only with this model? And is this choice ‘validation’ as they call it?

R. Thank you for your interesting observation. The validation is only proposed as an example to assess the validity of the methodology, only for a class of cities and not for all the three classes.

 

R405. The conclusion is a bit vague. What do they mean with ‘it’ in line 443. I would invite them to discuss who can use their model. What kinds of policy insights can come out of their model?

R. Thank you for your interesting observation.  The meaning of the sentence has been made clearer. Some discuss and implications are added in the previous paragraph “Results and discussion”.

Reviewer 5 Report

Comments and Suggestions for Authors

The manuscript shows an analysis on the influences of safety and energy expenditure parameters on cycling travel choice models. The idea of the manuscript is interesting but I do some concerns:

1- it is not clear what is the dependent variable (% Bike Share). I understood that the data were aggregated, but was the original variable binary (share or not share bike)?

2 - Can you show a histogram of this dependent variable  (% Bike Share) for checking for normality?

3 - The relationship between the variables is linear or non-linear? Please show some scatterplots.

4 - When a have read the title and the introduction, I was wondering if these new variables "cycle lane supply (CLS)" and "value of the energy expenditure (EE)" would be added to the models with other classical variables, as reported in the literature review. But the model has only these variables. I am not sure if this is the best way to show that they are influencing the % Bike Share.

5 - I am not convinced that this linear regression with 2 covariables and no intercept is give me new information on cycling travel choice. Why the intercept is zero? If the intercept is zero then the R2 is invalid....

Some minor review:

-in the entire text, please use "transport" (British) or transportation (American), but never both.

Comments on the Quality of English Language

The English is fine, but in the entire text, please use "transport" (British) or transportation (American), but never both.

Author Response

Reviewer 5

The manuscript shows an analysis on the influences of safety and energy expenditure parameters on cycling travel choice models. The idea of the manuscript is interesting but I do some concerns:

R501. It is not clear what is the dependent variable (% Bike Share). I understood that the data were aggregated, but was the original variable binary (share or not share bike)?

R. Thank you for your observation. Bike share is expressed as the number of cyclists that use bike for commuting over the total number of users that make a movement for commuting purposes with other modes of transport (not a binary variable).

 

R502. Can you show a histogram of this dependent variable (% Bike Share) for checking for normality?

R. Thank you for your observation. A histogram to check normality is proposed.

 

R503. The relationship between the variables is linear or non-linear? Please show some scatterplots.

R. Thank you for your observation. More graphs are added in order to explain the relationships between variables.

 

R504. When a have read the title and the introduction, I was wondering if these new variables "cycle lane supply (CLS)" and "value of the energy expenditure (EE)" would be added to the models with other classical variables, as reported in the literature review. But the model has only these variables. I am not sure if this is the best way to show that they are influencing the % Bike Share.

R. Thank you for your interesting observation. The authors agree. The other variables (as selected in Table 2) are added to the model.

 

R505. I am not convinced that this linear regression with 2 covariables and no intercept is give me new information on cycling travel choice. Why the intercept is zero? If the intercept is zero then the R2 is invalid....

R. Thank you for your interesting observation. The authors agree. The intercept is added to the models.

Some minor review:

R506. In the entire text, please use "transport" (British) or transportation (American), but never both.

R. Thank you for your interesting observation. In the whole paper only the “transport” term is used.

Round 2

Reviewer 4 Report

Comments and Suggestions for Authors

they processes my comments adequately

Author Response

Thank you again for your interesting comments. 

Reviewer 5 Report

Comments and Suggestions for Authors

The authors have answered all my points, but I feel that the lack of normality of the dependent variable (as expected because it is highly skewed to the right), can make the model missundertood. Indeed, the best model for this data is the beta regression, instead of the classical regression. But, the author can try to apply the log to the dependent variable in order to approach to normality.

The problem in use non-normal data in a classical regression is about the significance of the parameter estimates, since the p-value is under estimate, i.e, the variable is considered significant when in fact it is not.

So, I would like to check this with the authors. If my suspicious is confirmed, so the conclusion will change.

Author Response

The authors have answered all my points, but I feel that the lack of normality of the dependent variable (as expected because it is highly skewed to the right), can make the model misunderstood. Indeed, the best model for this data is the beta regression, instead of the classical regression. But, the author can try to apply the log to the dependent variable in order to approach to normality.

The problem in use non-normal data in a classical regression is about the significance of the parameter estimates, since the p-value is under estimate, i.e, the variable is considered significant when in fact it is not.

So, I would like to check this with the authors. If my suspicious is confirmed, so the conclusion will change.

 

Reply: Thank you for your constructive criticisms, for the interesting comments and suggestions for improvements: the authors have appreciated these ones and these comments have improved the scientific quality of the article. The log to the dependent variable is applied in order to approach normality: the models are a little bit changed, such as the results obtained, but, as a matter of fact, the distribution of the transformed dependent variable is normal as reported in the attached file.

Author Response File: Author Response.pdf

Round 3

Reviewer 5 Report

Comments and Suggestions for Authors

The authors made all requested review, and I feel the manuscript is ready to be accepted