Social Media Data Analytics to Enhance Sustainable Communications between Public Users and Providers in Weather Forecast Service Industry

Round 1

Reviewer 1 Report

The article is cognitively interesting. However, it raises doubts about the significance of the results. At the beginning, the question is, is the problem posed correctly? Can opinions in social media (tweets) improve communication between public users and providers in weather forecast service industry? Do they avoid errors in forecasts? The presented analyzes and results show what errors in forecasts and under what conditions trigger the most critical opinions of public users.
1. Introduction: It's very general. The importance of social media in market communication is too poorly shown. It is necessary to show - on the basis of literature - the importance of social media in public opinion polls. How to use tweets? What to do based on the analysis of tweets? How to change the broadcasted messages? What are the recipients sensitive to? What do they expect? How to change communication channels? Answers to these questions help to improve the communication process. This should be referred to in the theoretical framework.
2. Research project: does not raise any objections. It is definitely extensive and thoughtful.
3. Description of research methods: good. One could only clearly separate the description of the test methods from the description of the results (Section 2. Materials and method).
4. Description of the results: good. A bit too many quoted statements of the respondents. They do not contribute much to the content of the article.
5. Conclusions: there is no direct link between the conclusions and the results. It is a consequence of a complicated research problem. Therefore, the conclusions are of a general nature. On this basis, first of all, you need to describe how to change the weather forecast so that it contains fewer errors and causes less critical assessments.

I hope the authors will consider my comments and make corrections that will make the paper even better.

Author Response

The modifications to your opinion are as follows.

1. Introduction: It's very general. The importance of social media in market communication is too poorly shown. It is necessary to show - on the basis of literature - the importance of social media in public opinion polls. How to use tweets? What to do based on the analysis of tweets? How to change the broadcasted messages? What are the recipients sensitive to? What do they expect? How to change communication channels? Answers to these questions help to improve the communication process. This should be referred to in the theoretical framework.
☞ Thank you for the comment. We agree with what you have pointed out about supplementation of literature survey for social media research in a service sector. We augmented overall Section 1 part according to the complemented reference list. You can find the supplemented sentences in Section 1 on page 2-3 (line number 29-36, 55-66, 70-87).

2. Research project: does not raise any objections. It is definitely extensive and thoughtful.
☞ Thank you for your positive comment.

3. Description of research methods: good. One could only clearly separate the description of the test methods from the description of the results (Section 2. Materials and method).
☞ Thank you for the comment. We understand you are commenting that Table 1 and its related contents should be sent on a Section 3. This study focused on finding implications that could improve communication between meteorological community and service users by using association rule mining. As a result, we did not want to deal with the sentimental analysis methods and the sub-findings derived from the natural language analysis that have already been carried out in many studies. Therefore, we thought that it would be better to include Table 1 and its related information in Section 2 as the basic data characteristics before performing association rule mining analysis.

4. Description of the results: good. A bit too many quoted statements of the respondents. They do not contribute much to the content of the article.
☞ Thank you for the comment. Following your opinion, we deleted quotations from Section 3.1. We would like to leave the quotation in Section 3.2. to enhance the reader’s understanding about the exact meaning of the acquired association rules.

5. Conclusions: there is no direct link between the conclusions and the results. It is a consequence of a complicated research problem. Therefore, the conclusions are of a general nature. On this basis, first of all, you need to describe how to change the weather forecast so that it contains fewer errors and causes less critical assessments.
☞ Thank you for the comment. The line number 424~427 describes the implications derived from Rule A~D obtained by the association rule mining. And Rule G has drawn implication of line number 428~431. However, the mentioned parts did not clearly express the sources of the implications, which confused you. Thus, we added “see Rule A~D” in line number 427 and “see Rule G” in line number 431 to enhance readability.
Ultimately, in this study, we wanted to talk about the following possibility. Although the nature of weather forecasts makes it difficult to reduce overall errors, certain error types, such as Miss or False Alarm, can be corrected by adjusting the threshold of precipitation probability forecasts. Therefore, this study has identified the types of errors that users are sensitive to each season, and the results show the possibility that user satisfaction can be increased by enhancing Miss or False Alarm errors even though overall errors cannot be reduced. This is also mentioned in Section 4 in detail.

Reviewer 2 Report

Congratulations for the article. You have chosen a market niche. It is an interesting topic but I think is not very related to the topic conference.

First of all, please extend your state of the art and try to find in a creative way, a connection between researchers approach over the years.

Second you have sel-citations.

Maybe an econometric model it is required for the paper to be complete.

Author Response

The modifications to your opinion are as follows.

1. First of all, please extend your state of the art and try to find in a creative way, a connection between researchers approach over the years.
☞ Thank you for the comment. We agree with reviewer’s comment. We augmented overall Section 1 part with the complemented references including recent precedent studies. You can find the supplemented sentences in Section 1 on page 2-3 (line number 29-36, 55-66, 70-87).

2. Second you have sel-citations.
☞ Thank you for the comment. We removed self-citations (Reference 7 and 8 on the original manuscript).

3. Maybe an econometric model it is required for the paper to be complete.
☞ Thank you for your suggestion. We agree with your comment, and believe that the comment really points out the highlight of this study. However, the reality is that there is not enough time to carry out additional research methods you have proposed. So, we try to adopt your idea in future research.