Twitter Data Mining for the Diagnosis of Leaks in Drinking Water Distribution Networks
Round 1
Reviewer 1 Report
Research summary:
The topic is relevant and may be of interest to a broad range of the journal's readers. However, this reviewer has some major concerns about the paper.
Major Strengths: The major strengths of the research are:
- The topic is interesting
- The proposed approach has been properly designed and developed
Major Weaknesses: The major weaknesses of the research are:
- A real experimental evaluation is missing
- The structure and contents of the paper need to be improved.
Grammar and Readability:
The paper is well written and clear. I didn't find any typos.
Specific Comments: My specific comments concerning this manuscript are:
- The abstract does not highlight the specifics of the research or findings but contains too much background information. Some details of the research would be nice for example numbers addressing the sample, data, percentage improvement, etc.. Remove some of the background material and add some details of the research. Moreover, it is good to provide some specifics (e.g., sample size, dataset size, numbers from results, etc.).
- Although a novel combination might be allowed, it is necessary to highlight the contribution of such a combination from both methodological and empirical perspectives. Also, it is required to provide technical details of the proposed methods as much as possible and in-depth explanations on method selections.
- The innovation of the paper seems limited. The proposed method is a straightforward combination of existing techniques, which makes it less innovative. Also, more details of the proposed method should be provided.
- There needs to be an explicit research objective(s) and/or research question(s) stated, preferably as a separate section. This helps readers find out what the research is trying to address.
- There are several papers that have addressed similar problems, but it is necessary to further highlight the novelty between the proposed study and the related literature.
- Starting from the previous works, I suggest introducing a table to summarize the most recent works and to highlight the novelty of the proposed work.
- Please add more recent references. Certainly, there has been more recent (within the last two years) research on this topic published in information science and/or computer science outlets. An academic search on the topic (using keywords from the manuscript’s title) shows that there is recent work in this area. Therefore, authors must update their literature review. I suggest some recent works addressing similar problems: https://doi.org/10.1016/j.ipm.2022.103095, https://doi.org/10.1017/dmp.2020.347, https://doi.org/10.3390/foods11172695
- The reference list needs tidying up, as there are references missing items or formatting issues. Please be consistent with the formatting and use some standard formatting style.
- Please consider using a more convincing way to evaluate the proposed method.
Concluding Remarks:
The goal of the paper is interesting. I think that the paper could be improved with the considerations I reported in the review
Author Response
Reviewer 1
The topic is relevant and may be of interest to a broad range of the journal's readers. However, this reviewer has some major concerns about the paper.
Major Strengths: The major strengths of the research are:
- The topic is interesting
- The proposed approach has been properly designed and developed
Response: Many thanks to Reviewer 1 for this comment. It means a lot to us and encourages us to continue our research.
Major Weaknesses: The major weaknesses of the research are:
- A real experimental evaluation is missing
- The structure and contents of the paper need to be improved.
Response: Thank you also for this comment that prompted us to restructure the content of the article and to find a way to experimentally evaluate the methodology we propose.
 Grammar and Readability:
The paper is well written and clear. I didn't find any typos.
Response: We checked the document for grammatical and typographical errors that we may have missed. We corrected all the errors found.
 Specific Comments: My specific comments concerning this manuscript are:
 - The abstract does not highlight the specifics of the research or findings but contains too much background information. Some details of the research would be nice, for example numbers addressing the sample, data, percentage improvement, etc. Remove some of the background material and add some details of the research. Moreover, it is good to provide some specifics (e.g., sample size, dataset size, numbers from results, etc.). 
Response: The abstract was modified according to the suggestions of Reviewer 1. Background material was removed and replaced with specific information that was used to assess the methodology. The corrected abstract reads as follows:
“This article presents a methodology for using data from social networks, specifically from Twitter, to diagnose leaks in drinking water distribution networks. The methodology involves the collection of tweets from citizens reporting leaks, the extraction of information from the tweets and the processing of such an information to run the diagnosis. To demonstrate the viability of this methodology, 358 Twitter leak reports were collected and analyzed in Mexico City from May 1 to December 31, 2022. From these reports, leak density and probability were calculated, which are metrics that can be used to develop forecasting algorithms, identify root causes and program repairs. The calculated metrics were compared with those calculated through telephone reports provided by SACMEX, the entity that manages water in Mexico City. Results show that metrics obtained from Twitter and phone reports were highly comparable, indicating the usefulness and reliability of social media data for diagnosing leaks.”
-  Although a novel combination might be allowed, it is necessary to highlight the contribution of such a combination from both methodological and empirical perspectives. Also, it is required to provide technical details of the proposed methods as much as possible and in-depth explanations on method selections. 
Response: More technical details about the .methodology are given in Section 3. In fact, two detailed algorithms are presented in this section.
-  The innovation of the paper seems limited. The proposed method is a straightforward combination of existing techniques, which makes it less innovative. Also, more details of the proposed method should be provided.
Response: More details about the methodology are given in Section 3 to show that the main contribution of this article
- There needs to be an explicit research objective(s) and/or research question(s) stated, preferably as a separate section. This helps readers find out what the research is trying to address.
Response: The aim of the research is stated in Section 2 as follows after the presentation of the state of the art: “The aim of this research is to investigate the potential use of social media data, specifically Twitter, for detect, locate and evaluate leaks in drinking water distribution networks”.
- There are several papers that have addressed similar problems, but it is necessary to further highlight the novelty between the proposed study and the related literature. Starting from the previous works, I suggest introducing a table to summarize the most recent works and to highlight the novelty of the proposed work.
Response: We have included Table 1 in the article, which lists the more recent works on Twitter data mining. This table has four columns for ordering the information: the work reference, the objective of the work, the methodology used for developing the work and the application field.
 
 
- Please add more recent references. Certainly, there has been more recent (within the last two years) research on this topic published in information science and/or computer science outlets. An academic search on the topic (using keywords from the manuscript’s title) shows that there is recent work in this area. Therefore, authors must update their literature review. I suggest some recent works addressing similar problems: https://doi.org/10.1016/j.ipm.2022.103095, https://doi.org/10.1017/dmp.2020.347, https://doi.org/10.3390/foods11172695
Response: We have added the references suggested by the reviewer.
 - The reference list needs tidying up, as there are references missing items or formatting issues. Please be consistent with the formatting and use some standard formatting style. 
Response: We checked the references to correct those with the wrong format and complete those with missing information.
 - Please consider using a more convincing way to evaluate the proposed method.
Response: To evaluate the feasibility of using leak reports made through Twitter, we requested from SACMEX, the organization that manages water in Mexico City, all the reports made by citizens via telephone during the year 2022. These reports include the address of the leak and the date on which the reports were made. Three metrics were calculated from the reports made via telephone and Twitter: leak report density, leak report probability, and percentage of leak reports by municipality. These last two metrics are dimensionless and normalized with respect to the total number of reports, so they were used to compare the results obtained with telephone reports and reports via Twitter. Both types of results were similar with some discrepancies that deserve to be studied within a social and economic context.
 Concluding Remarks:
The goal of the paper is interesting. I think that the paper could be improved with the considerations I reported in the review
Response: Thank you so much for all the recommendations.
Reviewer 2 Report
The article is well written and in my opinion these are the points that improved
the manuscript.
1. Introduction needs to be revised with current reference.
2. I suggest authors should provide separate section for discussion and contribution of the result.
3. Limitation and future recommendation part should be separate from the conclusion.
4. Several abbreviations are used here, if possible please create an abbreviation table.
5. Author should elaborate results and discussion more clearly.
6 The steps of the Algorithm-1 are not clear (written in pseudo code) . Write the steps of the Algorith-1 more clearly
7 Please test the impact of the increase of important parameters in the sensitivity analysis.
8 There are many grammatical and typo errors throughout the document that should be corrected before this manuscript is published.
Author Response
Reviewer 2
- Introduction needs to be revised with current reference.
Response: We have included Table 1 in the article, which lists the more recent works on Twitter data mining. This table has four columns for ordering the information: the work reference, the objective of the work, the methodology used for developing the work and the application field.
- I suggest authors should provide separate section for discussion and contribution of the result.
Response: We have added an additional section called “Discussion”.
Limitation and future recommendation part should be separate from the conclusion.
Response: A section called “Discussion” was added before section “Conclusions”. In this additional section, we discuss the advantages and drawbacks of the proposed methodology, as well as future work.
Several abbreviations are used here, if possible, please create an abbreviation table.
Response: We have added a list of abbreviations before the references.
5. Author should elaborate results and discussion more clearly.
Response: The description of the results was improved by explaining in detail some metrics computed from the reports on Twitter and the reports via telephone. In shorts, we present in the results an evaluation of the feasibility of using leak reports made through Twitter. For this evaluation, we requested from SACMEX, the organization that manages water in Mexico City, all the reports made by citizens via telephone during the year 2022. These reports include the address of the leak and the date on which the reports were made. Three metrics were calculated from the reports made via telephone and Twitter: leak report density, leak report probability, and percentage of leak reports per municipality. These last two metrics are dimensionless and normalized with respect to the total number of reports, so they were used to compare the results obtained with telephone reports and reports via Twitter. Both types of results were similar with some discrepancies that deserve to be studied in future work within a social and economic context.
6 The steps of the Algorithm-1 are not clear (written in pseudo code). Write the steps of the Algorith-1 more clearly
Response: We modified the algorithm to make it clearer. Also, we added some lines for the calculation of leak metrics.
7 Please test the impact of the increase of important parameters in the sensitivity analysis.
Response: The sensitivity analysis is proposed as future work in Section 5, since for its realization different versions of the algorithms presented in this article will be run simultaneously for a long period (approximately one year). Each of these versions will have a modified parameter, for example, the search radius or keywords. The results obtained by the different versions of the algorithms will be analyzed to determine how the variations of the modified parameters affect the results.
8 There are many grammatical and typo errors throughout the document that should be corrected before this manuscript is published.
Response: We proofread the document to find grammar and typo errors. We corrected all the errors found.
Thank you so much for your recommentations that were helpful to improve our article.
Round 2
Reviewer 1 Report
Thanks to the authors for following the previous comments. The paper has been significantly improved and is ready for the publication.
Reviewer 2 Report
The authors have incorporated all the suggestions made by reviewers. The revised version is well written and I accept the paper in present form
