A Network-Based Analysis of a Worksite Canteen Dataset
, Marco Grassia 2, Alessandro Longheu 2, Michele Malgeri 2 and Giuseppe Mangioni 2Round 1
Reviewer 1 Report
This paper presents the analysis of a data set concerning with the chosen dishes in a multi company canteen. The main contribution of the paper lies in the network-based approach to analyze the dataset in addition to the classical statistical analysis that deeply is discussed throughout the paper. The focus of the paper is on the impact of people habits in eating during working and the impact of the proposed courses on the quality of life, however the technique used can be easily extended to other application fields. The paper highlights the importance of such approach in the case of large data sets since it allows the authors to perform a deeper analysis of the correlations among the different information contained in the dataset, the approach permits to investigate not only the correlation between couple of information but also permits to extract a graphic model that provides a very interesting view of the correlations among multiple axes. The case study is very interesting by itself and permits to validate the assumptions on the importance of a network-based approach. The work is well motivated and clearly presented and it can be accepted with some minor revision.- Substitute "#" with "Number" in the caption of figure 2
- Modify "Cat1" and "Cat2" with more explanatory labels
- Capitalize all captions.
- Check for any typos and linguistic inaccuracies.
Author Response
Please see the attachment
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Dear Authors,
thanks for this interesting investigation on canteen data.
I do have following questions and remarks:
1) I do like to introduction. It gives a nice overview and has the appropriate (number of) citations.
2) When introducing the data set it would be of interested from which country/region the data was taken from. The reader can guess from the authors list but that could lead to wrong assumptions
3) The average number of employees in line 92 is misleading as the distribution is not normal (or gaussian) and is not used later anyway.
4) Line 107: You group into 5 categories. It is not clear why not more or less categories? Is there a reference for having these 5 categories? Any other reason?
5) Line 102 vs. Line 125: "324 dishes in the data set" vs "317 different dishes were servered". Can you clarify this point?
6) Figure 3: I would present the data in form of a histogram, with frequency over nutrients and a binning of 5 and a range of [0,200]. The current graph implies some order of the dishes which is not existing.
7) There is always a large debate on whether to use graph technologies or simple (relational) data bases. In your case you investigate your data in form of a network. In my opinion your presented analysis does not benefit from using a network (graph). You could easily get to the same conclusion just using standard statistics and plots (histograms). E.g. compute number of dishes per person -> plot as histogram. Then also the total number would clear, while in you figure 5 the total number needs to be recalculated.
8) Figure 9 and Figure 10 are not referenced in the text. Either remove or include them to your text/analysis
9) The community investigation is missing the motivation. Do you expect any correlation? Maybe due to similar menu cards at different companies? What do we learn form the result. You conclusion is quite short about that.
10) There are some minor spell errors in the text, e.g. line 66 ", but it is quite ..."
11) Almost all figure captions need to be improved. Spelling and content are not appropriate.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Reviewer 3 Report
Authors are suggested to address the following issues.
- Title, please address:
(i) In general, the research topic should be focusing on a research topic, which can be evaluated by benchmark dataset. It seems that authors focus on a single dataset, instead of a generic approach. Please confirm if the network-based analysis could be applied to the application in canteen service.
(ii) Consider updating the title and specifying the technique and research area.
- Abstract, please address:
(i) It is too short. Please elaborate the key findings/results, as well as the improvement compared with existing works.
(ii) Authors have listed various assessment criteria and later on summarizing the contents of the paper as “a multi-company canteen service dataset and its network-based analysis is presented and first significant considerations, as well as future directions, are discussed”. The assessment criteria have not been included in the contents.
- Section 1, Introduction, please address:
(i) Paragraph 1, elaborate the necessity of the research topic “worksite canteens”. Also, it seems that student canteens share similar characteristics as worksite canteens.
(ii) Some of the contents belong to the discussion of related work, please move those contents into Section 2.
(iii) Summarize the contributions of the paper, preferably in point-form.
- Section 2, Related works, please address:
(i) Revise the contents in order to summarize the methodology, performance, and limitations of existing works.
(ii) Include more works (recent three years).
(iii) Explain the reason for the consideration of the tripartite network model.
- Section 3, Dataset description and representation, please address:
(i) Paragraph 1, share the reference for the dataset.
(ii) Authors have mentioned various existing works in Section 2, do authors use the same dataset as presented in existing works?
(iii) The format of Tables is not correct. Please refer to the journal’s template.
(iv) Column headings, avoid using abbreviations like “Co.id”, “#”, and “avg”.
(v) Elaboration should be made of Figure 1. Authors mentioned “tripartite network that allows us to connect companies (each with a different color), employees and meals”, please explain with the aid of Figure 1.
(vi) What are #1-11 in Figure 2?
(vii) Figure 3, what are xlabels and ylabels? Also, please use “.” instead of “,” for numerical value.
- It is suggested to improve the resolution for all figures. Some contents are blurred.
- Section 4, Dataset analysis, please address:
(i) Is the dataset scale-free or small-world network?
(ii) The rationales for using network analysis are not clearly explained. In food and nutritional sciences, there are other standard approaches to evaluate the food items.
- Please discuss the limitations of authors’ work and share future research directions.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Dear authors,
sorry for the late review. Thanks for the extensive reply on my comments.
I guess and I hope that the paper will lead to further publications and references, also including the discussion about the network approach for this data set.
Minor remark: Figure captions should be in general longer. A figure should be understandable solely with reading the caption without having the need to jump between the text and the figure.
Best regards
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
To facilitate review process, it is suggested to highlight the changes in the revised manuscript. Please refer to my follow-up comments:
Abstract, please address:
(i) It is too short. Please elaborate the key findings/results, as well as the improvement compared with existing works.
Authors’ reply: The abstract has been extended according to the reviewer’s suggestions.
Follow-up comment: This comment has not been fully addressed. Only one sentence is shared “Finally, the results prove that this kind of analysis can provide significant information that complements other traditional methodologies.”
Section 1, Introduction, please address:
(i) Paragraph 1, elaborate the necessity of the research topic “worksite canteens”. Also, it seems that student canteens share similar characteristics as worksite canteens.
Authors’ reply: The point has been clarified, see lines 16-18.
Follow-up comment: Reference is required for “In this work, we focus on worksite canteens but some of the results and the methodology can be applied to several others, like school canteens. The common characteristic is the presence of complex correlation among meals, as well as people and habits that usually arise in creation of communities.”
(iii) Summarize the contributions of the paper, preferably in point-form.
Authors’ reply: The introduction and the abstract has been partially elaborated as explained above. See lines 8-12, 51-59
Follow-up comment: The last contribution “to provide a different view of the dataset, represented as a tripartite network.” should specify the benefits and contributions when tripartite network approach is adopted.
Section 2, Related works, please address:
(i) Revise the contents in order to summarize the methodology, performance, and limitations of existing works.
Authors’ reply: The section has been partially revised, adding for each paper some further information about its achievements compared with our work.
Follow-up comment: Numerical results should be shared, like sections 3 and 4.
(ii) Include more works (recent three years).
Authors’ reply: We added some works according to this suggestion, see lines 99-110.
Follow-up comment: If this work is getting published (in 2021), the major portion of references in related works is not latest reference. Either authors adding more latest references or replacing the relatively old references with the latest references.
(iii) Explain the reason for the consideration of the tripartite network model.
Authors’ reply: We believe that a network representation of the dataset easily highlights relations among companies, people and courses through a tripartite network representation. From this, we can further elaborate the so-called network "projections", where the network among specific kind of nodes (e.g. people or dishes) is inferred. This allow to perform community analysis using standard network community detection algorithms.
Follow-up comment: Please ensure the comment has been addressed and included in the main text of the revised manuscript.
Section 3, Dataset description and representation, please address:
(i) Paragraph 1, share the reference for the dataset.
Authors’ reply: The company’s dataset is available only on request due to privacy concerns. We can share it with researchers that will sign a privacy agreements.
Follow-up comment: Please indicate in main text how the access can be requested.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Round 3
Reviewer 3 Report
Authors have addressed my previous comments.
I have noticed there are some minor comments.
(a) Line 189, correct “50K (49, 539) dishes” as “49,539 dishes”.
(b) Table 1, last column, it should be “Average number of meals”.
(c) Table 3, the column headings should be minimum and maximum, respectively.
(d) Figure 8, proper x-axis and y-axis should be given for each subfigure. Write the full name of “carbs”. Correct “KCal”.
(e) Figure 9, write the full name of “carbs”
Author Response
Dear reviewer,
all requests have been fulfilled.
Best regards
