Measuring Destination Image Using AI and Big Data: Kastoria’s Image on TripAdvisor
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsDear Authors,
The article "Measuring Destination Image Using AI and Big Data: Kastoria’s Image on TripAdvisor" presents a interesting exploration of how artificial intelligence (AI) and big data analytics can be used to understand and measure tourist destination images (TDI) as reflected in online travel review platforms. The research provides a solid and holistic approach to data analysis, which is well-suited to the objectives of the study.
A noteworthy aspect of the methodology is the analysis of color palettes derived from user-uploaded photos. By associating colors with specific emotions, the study introduces a creative and insightful dimension to understanding the emotional and aesthetic components of the TDI.
The use of user-generated content of TripAdvisor provides a robust dataset, and the research effectively extracts meaningful insights despite some limitations in the methodological tools which are mentioned. The visual tools, such as word clouds and sentiment graphs, enhance the presentation of the findings.
The results are well-presented and focus on significant themes such as seasonality, user engagement, and the emotional impact of visual content. The discussion ties these findings to practical implications for destination management organizations.
Author Response
We sincerely thank you for taking the time to review this manuscript and for your positive feedback. Your thoughtful evaluation is greatly appreciated.
Reviewer 2 Report
Comments and Suggestions for AuthorsYour paper is quite interesting and valuable for decision makers in tourism industry. You focused on Kastoria, Greece and through this paper you examined the tourist impressions of this historic city. You used a subset of TripAdvisor reviews about Kastoria that focus on its nature sights and historic attractions. You performed a thorough analysis of those selected reviews, where you wanted to find out e.g. when are the peak months of the tourist season and what attractions are more interesting for tourists from different parts of the world. Also, you performed an manual and automated analysis of the review photographs to examine what type of photos were most common and what are the most frequent objects that appear on them. Furthermore, you used AI tools to make a connections between the photos' predominant color and corresponding human emotions. Finally, you analyzed using AI tools the review texts to find out their overall sentiment and compare it with the ratings that the users entered with their reviews.
The language quality and the clarity of the text are very good, but I will recommend that you make some revisions before your interesting paper is considered for publication in this Journal.
Here is a short list of comments for your consideration. I hope that you further increase the quality of your paper by addressing those comments.
page 1 - line 7:
Please, add "(eWoM)" after first mentioning electronic word-of-mouth.
page 2 - line 80:
Please, avoid phrase "temporal tourists" and use just "tourists" instead.
page 2 - line 83:
Please, clarify what is "MMS" (I presume you meant social media).
page 3:
It would be interesting to add a comment (here or in the Discussion) on the differences/similarities between the word clouds for reviews in English, Greek and German.
page 3 - line 130:
Please, add a new reference to TextBlob (https://textblob.readthedocs.io/en/dev/) just as you did with TripAdvisor.
page 4 - lines 148 and 152:
Please, capitalize python to Python.
page 4 - lines 150:
Please, rephrase "and in addition" to "and additional libraries".
page 4 - line 153:
Please, add an explanation how and why did you associated 23 emotions with 10 main colors. It would be nice if you provided a table with the emotion-color pairs.
page 4 - line 164:
Please, consider changing "increase in critics" to "increase in critiques" or "increased criticism".
page 6:
For RQ3 you could add a disclaimer here that you inferred users nationality only based on the language they used to write the review. This is not always accurate because e.g. there are users from all over the world that write reviews in English.
Maybe it would be better to rephrase RQ3 on page 4 by avoiding mentioning nationalities, e.g. "Which attractions are most visited by tourists from different parts of the world?".
Also, you mentioned that you rejected the null hypothesis - for clarity you could state the null hypothesis and the alternative hypothesis for RQ3 even it can be easily inferred from line 207.
page 8:
Please, add clarification that in the first check you manually categorized 536 photos into 4 groups (containing no people, users themselves, friends, passers-by). The latter three groups also may be overlapping because it is not straightforward to decide without additional data who is a companion, passer-by or a user himself.
page 8 - line 236 and Table 3:
You mention 4240 descriptive words and in the Table 3 you mention the total number of object views as 4248. Should these two numbers be the same? Please, check and fix if necessary.
page 9 - Figure 5:
Please, add an explanation for this figure and what can a reader learn from it. E.g., if the majority of palletes are gray then the photo is dull and boring? It would also be nice to provide the actual photo in question if it is possible.
page 10 - line 303-304:
Please, consider adding a comment why this is the case - privacy reasons or emphasis on the places/attractions.
page 10:
You could expand the discussion with the color-emotion connections that you made, e.g. you can state what were the predominant colors in the 536 chosen photos.
page 10:
You could also add further comments about the discrepancies between the user review ratings and the polarity scores from the review texts. It seems that because of the language nuances it is very challenging for AI tools to decide if a review is good or very good, bad or very bad and most importantly neutral or good/bad. However, 92% reviews are rated by users as good or very good, and around 70% of reviews are categorized by the AI tool as good or very good which can be considered as acceptable.
page 11:
Please, add an explanation how this analysis of tourists' impressions of Kastoria, Greece can be applied to different places and tourist attractions around the world. You should emphasize how this research approach can be generalized for analysis of other tourist attractions around the world and for comparison between different tourist attractions.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsWith version 2 you have increased the quality of your very interesting paper. I commend you for making all the improvements and addressing all of the comments appropriately. The paper is now expanded with interesting additional details about emotions conveyed by the colors in photos and deeper discussions about Kastoria's tourism image.
I will recommend that your paper is accepted for publication in this respected Journal.