Citywide Metro-to-Bus Transfer Behavior Identification Based on Combined Data from Smart Cards and GPS

Round 1

Reviewer 1 Report

Please check some typo errors, like in line 64, you mistyped work with word etc.

Please explain more the Table 3, the dynamic model and survey data, what this means for your study?

 

Author Response

Question 1: Please check some typo errors, like in line 64, you mistyped work with word etc. 


 ANS: Thank you for your corrections. Based on reviewer’s comments, we have carefully corrected all typos and grammatical mistakes by a rigorously professional method. Please check the resubmitted manuscript with marked tracks.

Question 2: Please explain more the Table 3, the dynamic model and survey data, what this means for your study?

ANS: Thank you for your comments. We have revised it in the resubmitted manuscript, and explained more about the Table 3. Please check Line 411-420 on page 12 of the re-submitted manuscript. In order to conveniently check the revised descriptions for the reviewer, the specific revisions in resubmitted manuscript are cited below:

       “As shown in Table 3, the columns of survey data are the field occurring metro-to-bus statistic transfer data of partial passengers sampled by the 8 local volunteers, and the other columns represent the statistic results of all transferred passengers recognized by dynamic model in this study (Table 2) and static model with the one-size-fits-all criterion elapsed time threshold of 30 min, respectively. The row data of volumes means the total number of transfer trips at the specific time period and metro station, and the average or variance represent the statistic value of all trips, respectively. Notably, the surveyed transfer volume is much smaller than the two others because the volunteers can only track the part of transfer passengers in practice. Therefore, our focus is to test the statistic gap among three data source, such as average, variable, and significance level.”

Author Response File: Author Response.docx

Reviewer 2 Report

THe paper is well written it is really easy to follow. I like it.

Author Response

Question 1: The paper is well written it is really easy to follow. I like it.


Response 1: We are appreciated for your comments. Based on the other reviewers, we have rigorously revised our manuscript. Please check the resubmitted version.

Author Response File: Author Response.docx

Reviewer 3 Report

 

 Citywide metro-to-bus transfer behavior – review comments

 

Suggestion: remove the last paragraph from the abstract, the one that begins with “Notably (…) 30min.”

In the introduction it is important to specify that metro refers to a subway system.

p1, l34 – the expression “cheap and convenient way” appears to be somewhat condescending; this could be re-written for a more neutral tone.

p1, l43 – questionnaire data being expensive and time-consuming; however, wasn’t this part of the methodology employed in this study?

p2, l71 – safety and the transfer distance; clarify what kind of safety it refers to (e.g. crossing the street, lack of lights, dangerous neighborhood?)

p3, ls107-108 – could a % (or a set of examples) be provided of how many cities/systems do not require swiping before alighting?

p3, l143 – time of the first passenger, is likely to be: boarding time of the passenger first swiping

p3, l144 – record time is likely to be: recorded time

p3, l144 – when passengers exit the ticket gate, would this be: when passengers pass through the ticket gate?

p5, ls177-180 – assuming that these scenarios refer to 2019, then March 17 was Sunday, not is Sunday;

p5, ls190-191 – is this finding consistent with those of similar systems?

p6, l257 – doesn’t this depend on (and is influenced by) the frequency of service and the service headways? (peak-hours shorter headways vs. non-peak hours: longer)

p9, ls336-337 – isn’t this an unwarranted statement? Since again it may depend on the frequency of service and headways.

ps9-10 – having volunteers follow passengers would enable them to also collect basic demographic data (gender, estimated age, walking speed, pauses for coffee or going to the restroom, etc.)

p13, l433 – more accurately and reliably (could it be: has much more accuracy and reliability)

p15, l485 – Figure 10 provides an overview?

Shouldn’t the references have abbreviated journals’ names instead?

Transportation planning and technology – ought to be Transportation Planning and Technology?

Author Response

Question 1: Suggestion: remove the last paragraph from the abstract, the one that begins with “Notably (…) 30min.”


ANS: Thank you for your suggestions. We have removed it in the re-submitted manuscript.

Question 2: In the introduction it is important to specify that metro refers to a subway system.

ANS: Thank you for your comments. In this study, we did not distinguish the metro and subway system because metro, tube, subway, or underground railway system are all names of rail systems operating underground in different cities of the world according to the website of https://www.worldatlas.com/articles/what-is-the-difference-between-a-metro-a-subway-and-an-underground.html . The different regions have different preferences when it comes to choice of names. In order to strictly explain our research issue, we have revised the manuscript in Lines 28-29. The specific revision in resubmitted manuscript are cited below:

“The urban metro or subway system, one of highest capacity traffic mode in the Mass Rapid Transit (MRT), can provide the reliable, fast and affordable service for the medium-long distance travelers.”

Question 3: p1, l34 – the expression “cheap and convenient way” appears to be somewhat condescending; this could be re-written for a more neutral tone.

ANS: Thank you for your comments. We have revised them with the neutral tone, and please check Line 31 on Page 1 of the re-submitted manuscript.

Question 4: p1, l43 – questionnaire data being expensive and time-consuming; however, wasn’t this part of the methodology employed in this study?

ANS: Thank you for your comments. As known, it’s very expensive and unfair to collect traffic data via questionnaire. In fact, this paper developed a fast data fusion method for recognizing metro-to-bus transfer trips with only using the automatically collected actual operation data from smart card system and GPS system in the city of Shenzhen, China. Meanwhile, to verify the effectiveness and reliability of the proposed method, the authors recruited a sample of 8 local volunteers to collect the occurring passengers’ transfer trips at four metro stations in Shenzhen. The surveyed method is presented in Section 5.1.1.

Question 5: p2, l71 – safety and the transfer distance; clarify what kind of safety it refers to (e.g. crossing the street, lack of lights, dangerous neighborhood?)

ANS: Thank you for your suggestions. Based on the investigated results by Cherry and Townsend (2012), the greatest impacted factors are safety from crime and the distance between metro exits and bus stops [6]. We have further clarified them, and please check Lines 66-69 on Page 2 of the re-submitted manuscript.

Question 6: p3, ls107-108 – could a % (or a set of examples) be provided of how many cities/systems do not require swiping before alighting?

ANS: Thank you for your suggestion. To the best of our knowledge, no released data refers to how many cities have opened the smart card system to collect boarding or alighting information up to now. alternative, we could only add some cities (London, Santiago, Beijing, and Shenzhen) to demonstrate they do not record the alighting data. We have revised them in the resubmitted manuscript. Please check Lines 106-107 on Page 3 of the re-submitted manuscript.

Question 7: p3, l143 – time of the first passenger, is likely to be: boarding time of the passenger first swiping

ANS: Thank you for your correct. As the reviewer’s comment, we have revised them, please check Lines 140-142 on Page 3 of the re-submitted manuscript.

Question 8: p3, l144 – record time is likely to be: recorded time

ANS: Thank you for your correct. As the reviewer’s comment, we have revised them, please check Lines 140-142 on Page 3 of the re-submitted manuscript.

Question 9: p3, l144 – when passengers exit the ticket gate, would this be: when passengers pass through the ticket gate?

ANS: Thank you for your correct. As the reviewer’s comment, we have revised them, please check Lines 140-142 on Page 3 of the re-submitted manuscript.

Question 10: p5, ls177-180 – assuming that these scenarios refer to 2019, then March 17 was Sunday, not is Sunday;

ANS: Thank you for your correct. We have revised them, and please check Lines 176-177 on Page 4 of the re-submitted manuscript.

Question 11: p5, ls190-191 – is this finding consistent with those of similar systems?

ANS: Thank you for your comment. Yes, it is. We found that the average transfer time and volumes per hour are very approximate from 7:00 to 21:00 on weekdays in Figures 1 and 2, which is consistent with those of Zhao et al. (2019), who reported that the metro-to-bus transfer is relatively stable throughout the week. We have added them, and please check Line 182 on Page 5 of the re-submitted manuscript.

“Zhao, D.; Wang, W.; Li, C.; Ji, Y.; Hu, X.; Wang, W. Recognizing metro-bus transfers from smart card data. Transportation Planning and Technology. 2019, 42(1), pp.70-83.”

Question 12: p6, l257 – doesn’t this depend on (and is influenced by) the frequency of service and the service headways? (peak-hours shorter headways vs. non-peak hours: longer)

ANS: Thank you for your correct. Yes, the passenger's transfer time is influenced by many factors, such as transfer distance, bus frequency, arrival time, bus overload. We have revised them, and please check Line 247 on Page 6 of the re-submitted manuscript.

In details, the traditional static methods of one-size-fits-all criterion do not consider the station-based difference and departure time (peak-hours or non-peak hours). In Section 3, we found that the average transfer time and volumes are significantly fluctuated over different time periods. Therefore, the one-size-fits-all criterion is not suitable for each station or all time period in practice. To address this problem, this study divided nine time periods including three kinds of time of day (am-peak, pm-peak, and off-peak) from three kinds of day of week (weekdays, weekends, and festivals). In Rule 2, we firstly set a static threshold of 40 min for screening, where Jang et al. found that almost transfer trips have a transfer time of less than 40 min). And then, the 95^th percentile of filtered passengers’ travel time is regarded as the metro-to-bus elapsed time threshold for the specific time period.

“Jang, W. Travel time and transfer analysis using transit smart card data. Transportation Research Record: Journal of the Transportation Research Board. 2010, 2144.1, pp.142-149.”

 

Question 13: p9, ls336-337 – isn’t this an unwarranted statement? Since again it may depend on the frequency of service and headways.

ANS: Thank you for your correct. Yes, it also depends on the frequency of service and headways. However, in the third chapter, we found the elapsed time thresholds should constant at the same time period (this study divided nine time periods) for the day of week at a specific station because the previous average transfer time and volumes are similar, such as peak-hours on weekdays in Figure 1. It is consistent with that of Zhao et al., who find that the metro-to-bus transfer is relatively stable throughout the week.

“Zhao, D.; Wang, W.; Li, C.; Ji, Y.; Hu, X.; Wang, W. Recognizing metro-bus transfers from smart card data. Transportation Planning and Technology. 2019, 42(1), pp.70-83.”

 

Question 14: ps9-10 – having volunteers follow passengers would enable them to also collect basic demographic data (gender, estimated age, walking speed, pauses for coffee or going to the restroom, etc.)

ANS: No. This study focuses on investigating the number of transfer passengers and related their transfer travel time from metro station to bus station, and so that volunteers did not record the demographic data. The authors think these kinds of data is much more suitable to analyse the transfer behaviour after transfer identification. Following the reviewer’s concerns, we will continue to study the impacted factors of transfer behaviour and present it in the future work. Please check it the re-submitted manuscript in Line 544.

Question 15: p13, l433 – more accurately and reliably (could it be: has much more accuracy and reliability)

ANS: Thank you for your correct. We have revised them, and please check Lines 442-443 on Page 13 of the re-submitted manuscript.

Question 16: p15, l485 – Figure 10 provides an overview?

ANS: Thank you for your correct. Yes, Figure 10 has summarized the distribution of metro-to-bus transfer elapsed time threshold estimated by the proposed method in this paper under the scenario S1, which includes 131 metro stations belonging to metro lines 1-5 in Shenzhen. We have revised it, and please check Lines 493 on Page 15 of the re-submitted the manuscript.

Question 17: Shouldn’t the references have abbreviated journals’ names instead?

ANS: Thank you for your correct. We have revised all references with the full journal name in the resubmitted manuscript.

Question 18: Transportation planning and technology – ought to be Transportation Planning and Technology?

ANS: Thank you for your correct. We have revised them in the resubmitted manuscript.

Author Response File: Author Response.docx