Abstract

The study investigated the comments on marriage posts of Pakistani actors to analyze the interplay between negative and positive user-generated opinions. Utilizing tools such as MeaningCloud a comprehensive analysis was carried out on 2000 diverse sets of comments from the marriage posts of 10 Pakistani celebrities. The positive comments were loaded with expressions like 'MashAllah', and 'Cute Couple' to convey admiration for the newly married. However, the findings also reveal the presence of negative sentiments which also emphasize the complexities that arise when public opinion is shaped by conservative societies with strong religious influence. Negative comments are loaded with expressions like 'mujra’ for couple dance, and dislikeness for actors' life choices with words such as 'buri(bad)'.  Thus, it can be construed that on one hand, the positive comments reflect joy and appreciation for achieving a personal milestone like marriage, on the other negative comments indicate adherence to religious and cultural norms.

Key Words

Sentiment Analysis, Pakistani Actors, Instagram, Marriage Posts, Meaning Cloud

Introduction

The rise of social media has significantly impacted the way we communicate with each other. Social media allows us to create a persona of oneself which is easily available to others. Different social media platforms like Twitter, Facebook, and Instagram are popular worldwide for their photo and video-sharing abilities and allowing people to share their personal moments with each other however despite the utility of these platforms they are also known for hot drink online harassment and bullying. This research in disregard aims to carry out a sentiment analysis of selected posts of Pakistani actors on Instagram. 

There has been an exponential growth in the popularity of Pakistani actors on Instagram with many celebrities amassing millions of followers worldwide. These actors use the platform in Instagram to share their life events and happenings with their followers such as traveling experiences, shopping experiences, daily life happenings, and the stories of engagements and marriages. Consequently, these posts drag hundreds of comments from users who are eager to express their views about these celebrities. While there are many comments some are positive and some are negative with the spread of hate and even found to be harassing. The idea of anonymity and the perceived lack of consequences have exacerbated this problem making it an important idea to understand the sentiments and comments of different users to identify the instances of cyberbullying. 

To conduct this research, I have selected Instagram comments from popular Pakistani actors from Instagram to carry out a sentiment analysis of the comments associated with their marriage posts. A comprehensive data set of 4000 comments from the Instagram handles of different celebrities has been pre-processed by removing irrelevant information to carry out the process of sentiment analysis.

Sentiment analysis also called opinion mining is a subfield of NLP Sentiments behind our generated text. In this study, I'm going to apply sentiment analysis to the user comments on Instagram posts of Pakistani actors to contribute to the understanding of user behavior on Instagram posts, particularly in the context of celebrity marriage and discussions with surround them. One of the primary objectives of this study is to identify if there is any presence of cyberbullying on Instagram posts of selected Pakistani actors as cyberbullying is not only detrimental to emotional and mental health it also contributes to the production of a toxic environment on the social media platform. 

Literature Review

Sentiment analysis, in simple terms, is an application of various concepts in linguistics and artificial intelligence to understand the emotions behind peoples' speech, written and verbal. It is actually a sub-field of natural language processing (NLP) and is used to identify emotions behind words spoken by people (Liu, 2012). The changing dynamics of the world and the advent and proliferation of social media platforms have yielded a lot of data in terms of public opinion and comments. This has widened and necessitated the use of sentiment analysis to understand the feelings and attitudes behind words (Cambria et al., 2013).

As mentioned earlier, the generation and availability of data because of the proliferation of social media platforms have provided the researchers with ample opportunities to apply sentiment analysis. Data from various fields such as politics (Tumasjan et al., 2010), the consumer industry (Jansen et al., 2009), and the entertainment world (Kim et al., 2016) have been analyzed by using the framework of sentiment analysis. All these studies paved the way for this research, where researchers have applied sentiment analysis to understand and ‘quantify’ the emotions behind peoples’ comments on Instagram posts of celebrities’’ marriage ceremonies.

There have also been studies conducted on the content of social media using the framework of sentiment analysis. These studies have focused on content from blogs, forums Twitter, etc. (Pak & Paroubek, 2010). These studies did the groundwork for the furtherance of the application of sentiment analysis on various kinds of data, including the data for this research. The extensive growth of user-generated data has given many opportunities to researchers for the study of sentiments behind discourse (Agarwal et al., 2011).

The focus of the researchers has expanded to other social media platforms in the last few years. These platforms include Instagram, Facebook, and YouTube, to mention a few, and provide a wide range of data for sentiment analysis (Loria et al., 2014). The framework has itself gone through various developments, like the changes brought about by machine learning algorithms (Go et al., 2009), approaches involving lexicography and lexicon-based approaches (Taboada et al., 2011), and deep learning models (Mikolov et al., 2013).




Instagram as a Social Media Platform for Sentiment Analysis

The uniqueness of Instagram lies in its visually driven setting. However, this uniqueness also poses many challenges for the application of sentiment analysis. For researchers, there is a unique opportunity to analyze the text as well as the visuals shared by people and the comments and responses generated thereafter. Many researchers have recognized the uniqueness of the platform in providing this opportunity to the researchers (Bakhshi et al., 2014). Not only does it provide the opportunity for the analysis of the content shared by people, but also the response generated in the form of follows, likes, and comments by the people (De Choudhury et al., 2016). The researchers examine these interactions in order to better understand the various forms of public opinion and the way it spreads among social media users.

Since Instagram has become one of the most preferred platforms by celebrities to share their content, researchers have also more focused on the content shared on this platform. Several studies have been conducted to understand and lay bare how the posts shared by celebrities propagate and advertise various things and fashions for the common user (Hu et al., 2014).

All of this has paved the way for the focus of the present study on the Instagram posts of the marriage ceremonies of various Pakistani celebrities/actors and the comments those posts generate. The study is particularly focused on the comments made by people on those posts.




Applications of Sentiment Analysis in the Entertainment Industry

The application of sentiment analysis in the entertainment industry has been widespread for various understandable reasons (Dhar & Varshney, 2013). Researchers have conducted studies to predict the success of various box office releases (Asur & Huberman, 2010), the impacts of scandals on the lives of celebrities (Kwon et al., 2018), and how ardently people follow their favorite personalities (Mariconti et al., 2020).

All of these studies bring forth the possible benefits that can be taken from the application of sentiment analysis in the entertainment industry. The stakeholders involved, for example, actors, producers, directors, etc can make a lot of changes to their projects based on the data gleaned from the relevant research. This study, with its focus on the comments made by people in response to the posts shared by the actors' marriage ceremonies, can yield a lot of data and benefits for the entertainment industry.

It should be mentioned here that sentiment analysis is not only limited to the generation of this kind of data and predicting the success and failures of movies etcetera, it has also been employed to understand and investigate other serious issues of society. For example, there are various studies conducted with the framework of sentiment analysis which examine the reception and representation of various cultural norms and values, (Jiang et al., 2019), the situation of gender equality in the film and TV industry (Carter et al., 2018) and influences of popular music (Peterson & Berger, 1996). It is in this context that the present study has been carried out to understand the interplay of cultural values and public sentiment in the context of the entertainment industry of Pakistan.

Since sentiment analysis is capable of identifying various factors that influence the perception of the audience and multiple reasons for their engagement with various social media posts, it can help content generators evaluate and modify their strategies for effective communication and positive responses (Ruths & Pfeffer, 2014). In this case, this study, by analyzing the reaction to various Instagram posts of Pakistani actors, may enable the entertainment industry as a whole and the actors to understand the various positives and negatives of their posts and how they can improve their representation, communication, and acceptance with people. Not only that, but it can also be important in understanding the cross-cultural differences in the reaction of people to various kinds of entertainment (Li et al., 2018). The focus of this study is study of the response of the people to the posts of various Pakistani actors, and it may reveal how people from various backgrounds react to the posts of actors' marriages, their likes, and dislikes too, and the role that the cultural context plays in the formation of responses.

Cultural and Societal Aspects in Sentiment Analysis

Sentiment analysis gives a very good opportunity to study and analyze public sentiments, their context and background, and the reason for those sentiments (Balasubramanyan et al., 2012). Already, research has been conducted to understand the regionality of certain responses, for example, Abdulla et al. (2013) carried out a study on the Middle East, and (Chen et al., 2018) on Asia. The role of language variations, cultural values, norms, etc has been highlighted through these studies.

In the context of the entertainment industry in Pakistan, it is essential to understand various factors behind people's engagement with celebrities through social media and how they respond, particularly in the context of posts of the actors' marriages.

Methodology

This research is quantitative in nature and for text-mining the researcher has used the techniques of text clustering, topic extraction, text categorization, and sentiment analysis. For the analysis, the researchers have used MeaningCloud software, a powerful tool to extract meaningful knowledge from all types of unstructured content allowing a comprehensive analysis of various classifications of texts. MeaningCloud API uses a semantic approach based on advanced NLP to comprehend the human's natural language, allowing machines to interpret elements of human communication. For the analysis, a sample of 4000 comments was chosen out of which the researchers randomly selected 2000 comments for sentiment analysis.  To clean the data; the researchers used TextFilter and TextFixer. 

Analysis and Discussion

The analysis was used using distinct text-mining procedures. The first procedure is topic analysis for which the topics were extracted with the help of MeaningCloud. The researchers have found the 10 most mentioned topics in the comments of Instagram followers of Pakistani actors.




 

Table 1

Ten Most Mentioned Topics

Positive/Negative Words	Frequencies	Percentages
MashaAllah	409	10%
Happy	189	5%
Cute	75	2%
Congratulations	300	8%
Love	229	6%
Indian	83	3%
Lanat+beghairat	20	1%
Shame	20	1%
Fazool/bekar	15	0.3
Buri	25	1%

 

Analysis of Deep Categorization

Figure 1

Excerpt of Deep categorization from MeaningCloud Sentiment Analysis



There are various topics and themes present in the comments. These topics and themes can be arranged into certain categories, and when these categories are linked with the overall sentiment analysis, further insights can be had. 

1. Religion & Spirituality (Islam): the use of religion seems two-pronged: while some positive comments use religion for blessing the couples and their marriages, showing that people commenting do not consider these things irreligious, and they wish for the protection of those people against evil. Comments like "Masha Allah" and "Allah jj apko hamesha khush rakhe ameen" are expressions of that goodwill. There are, however, on the other hand, people who think these activities are irreligious and against the teachings of Islam and the Quran. They also highlight the national sentiment by alluding to the religiosity in the name of the country, Pakistan, the land of the pure.

2. Style & Fashion: although in a small number, there are comments which explicitly mention the good fashion sense and style of the people posting the content.

3. Personal Celebrations and Life Events (Wedding): marriages are considered the most important events in the life of people, especially here in Pakistan. They have also this religious side to them. So, both culturally and religiously it is considered needed to celebrate them. Therefore, there are some comments congratulating the parties involved. 

4. Family and Relationships: marriages are especially noteworthy for the involvement of family and relationships. There are many comments highlighting this side of the functions the pictures of which are posted on the social media platform.

5. Wedding Attire: one comment criticizes the choice of lehnga as the wedding attire for the reason that it was poorly designed. The comment can be interpreted as negative, pointing out the subversion of people's expectations with traditional wedding attire.

6. Careers (Pakistani celebrities/ actresses): A somewhat vague comment, there is not much that can be made out of it. It mentions the actors and actresses involved in the pictures but does not clearly articulate the sentiment.

In short, the deep categorization analysis reinforces the comments made above. It shows that seen vis-à-vis themes and topics extracted out of the comments, it can be said that there are some positive comments and some negative comments. While the positive comments celebrate the match-making and wish the parties involved well, the negative comments generally approbate the people involved. The reason for criticism comes from various factors like religion, culture, and personal liking and disliking. 

 




Analysis for Positive Comments

Figure 2

Excerpt of Positive comments from MeaningCloud Sentiment Analysis



The sentiment analysis of the comments by Meaning Cloud shows them to be positive. Meaning Cloud has marked all the comments to be non-ironic, which means that the sentiments expressed by people are genuine and there is no irony in them. Most of the comments are also labeled with the word ''AGREEMENT'" which shows them to be positive and favorable to the general message of the posts.

After the ÁGREEMENT', the next marker for the comments is a division between 'subjective'' and ''objective'". Subjective comments normally involve personal feelings and opinions while objective comments are fact-based, which can also be called neutral. In the comments fed to Meaning Cloud, both types of comments are there. Comments like 'sweet couple’ and ‘many many congratulations my sweet sis'' are subjective comments while others are objective.

Figure 3

Excerpt of Positive comments from MeaningCloud Sentiment Analysis



On the score chart of P to P+ (where P is positive while P+ is strongly positive), many comments are P+. these comments include ''Yaar very emotional'' and 'Wow that is awesome'' show more heightened sentiment than denoted by the P comments. In sum total, the comments indicate a strong approval and likeness for the content of the post, and concurrence with the cultural value and ethics of the comments. Since a good number of comments include words like 'beautiful sweet cute and awesome, this indicates that the overall sentiments are positive and the content of the posts is being received by the people favorably. Some admiration words like Mashallah also indicate the cultural and religious context of the posts. They show that the public feels a cultural and religious affinity with the people posting that content and genuinely wishes their protection and blessing upon them.

There are other words that use the word couple and sentiment attached to that, indicating that people understand and value the romance involved in the relationship. Since the posts are about the marriage ceremonies or marriage-related ceremonies of actors, it means that the people generally approve of the matchmaking and are happy for the couples involved. The word Mashallah in this context also reveals that the marriage or engagement needs to be protected from evil influence.

Apart from the emotional value of the comments and their general approval, there are also cases of cross-border enthusiasm and approval. For example, the comment ''Hey, fan of you from Iran'' shows that the audience also involves some cross-border or international players as well. This shows the wide-ranging appeal of the content posted and its cultural approval.

 




Analysis for Negative Comments

Figure 4

Excerpt of negative comments from MeaningCloud Sentiment Analysis

This excerpt from Meaning Cloud provides a sentiment analysis of the comments the tone of which is generally negative, disapproving, disagreeing, or abrogatory. These comments show peoples' disapproval or dissatisfaction with the content of the posts. There might be many reasons behind it, from personal dislike to religious, cultural, or class-based reasons. There is a division between the comments on account of their being subjective or objective. Like the previous marker, they are also categorized as 'non-ironic'' which shows that there is no duplicity in them. The disapproval or dissatisfaction expressed in them is genuine.

Figure 5

Excerpt of negative comments from MeaningCloud Sentiment Analysis



The criticism of the comments ranges from cultural to religious to national. There are also some hints of personal disliking. The comments involving the words 'mujra’’ and ‘naach gana culture'' are not only a cultural reprobation but also religiously motivated. Muslims believe that these kinds of things are against the teachings of the Quran, and it is their duty to remind people of the teachings of God. Since the content being analyzed involves the mentioning and celebrating of engagement/ wedding, there is disapproval of dancing at weddings. One comment reads/ suggests that after such banal expressions of vulgarity, it was high time to change the name of the country from Pakistan (the land of having strong religious undertones) to something else.

The comments involving the lack of coordination in the dance steps show some sort of personal dislike. Otherwise, there is not much reason to mention it. Although it can mean that the person making the comment did not like the quality of the performance in the post, still it seems to have been spurred by personal disliking. 

Some comments also highlight the uselessness of such activities, especially their advertisement on social media. Comments like ''hate such functions'' and ''fazool tareen'' highlight the general irrelevance and uselessness of such comments. In a country where nearly 40 percent of the people live under the poverty line, the advertisement of such lavish functions only throws salt on people's wounds. They express themselves through the comments. That is why there are also some comments that label these posts and the people in them as mere show-offs. They make the point that people involved in these posts are attention-seekers. The words hinting at peoples' immodesty have also been used, which show the religious as well as cultural disapproval of the content.

There are also comments hinting at the fact that these people are copy-catting the values and cultures of other people- most probably Western and Indian- and preferring the latter to their own culture. These comments are inspired by nationalistic sentiments as well. Apart from that it also suggests that people think that these kinds of things are not original. One comment saying that some particular event being advertised feels more like a ‘’walima’’ than a ‘mehndi’’ can suggest two things: one more obvious one is thematic inaccuracy. The second aspect, however, might have to do with the lavishness of the function. While walimas are generally large gatherings, mehndi is considered to be a smaller function. Declaring a mehndi function more like a walima might have to do with the spending: the people having the functions are spending on mehndi what is normally spent on a walima.

The overall impression of sentiment analysis of these comments is negative. The comments use various methods/ pretexts for disapproving the performance of those in the function. The approbation involves many reasons, as mentioned earlier, like personal disliking, affront to religious teachings, and against own culture.

Discussion

Sentiment analysis has become hugely important for decoding and understanding the emotions behind words. Research has shown that sentiment analysis can provide reliable data and insights into various social aspects. Its influence ranges from political opinions to the world of movie making, health, and people's preferences for certain products and services. This discussion analyzed those sentiments expressed in comments by people from different backgrounds. These comments are made in response to various posts by Pakistani actors/ actresses about their wedding ceremonies. 

In the comments, it was found that a lot of people were very positive in receiving the overall message of the posts. They not only identified with the sentiment but also expressed their well wishes for the involved. This shows that the marriage posts of actors generally produce positive responses among their followers and social media users. People's comments were motivated by shared cultural values, religious importance, and general positivity involved in marital festivities. It also involved celebrating, matchmaking, and love marriages. 

It is also important to highlight the cultural context of these comments. Marriage has very strong cultural and religious associations. In one of the injunctions of the Holy Prophet PBUP marriage has been equated with half of the faith. So, there is a celebration of marriage religiously. Secondly, even as a cultural expression, it is connected with the agrarian culture of the country where marriages have traditionally been a symbol of union and growth. For the actors to openly celebrate their marriages stroke these religious and cultural sentiments of people. It shows people that despite the actors being from a different league altogether, they still adhere to the religion and cultural sentiments of the people. That is why there are positive comments. This reinforces the idea of the previous research that cultural affinities invoke positive sentiments among online users of social media platforms (Balahur et al., 2011).

The context for the negative comments is almost the same. Some ultra-conservative people think that these kinds of dancing parties and show-offs, with mingling of opposite sexes, are against the general spirit of Islam. They also think that in a country gained in the name of Islam, these things should be discouraged. Apart from that there is also a class- angle to the whole argument. The expression of richness and wealth of these parties is directly contravening the lifestyles of many poor people in the country. Once again, this finding is consistent with the previous finding where it is concluded that in conservative societies religion plays a fundamental role in shaping people's practices and responses. (Boudad et al., 2018).

One of the challenges the sentiment analysis faces is the identification of the subjective expressions as these kinds of expressions are not transparent enough. They are often very much context-dependent, among other things. Several subjective expressions were recorded, however, highlighting people's personal sense and liking for style and aesthetics. These subjective expressions largely contribute to the heightening of like or dislike and make for strong responses.

Conclusion

The sentiment analysis carried out on the comments of the people on the wedding posts of various Pakistani actors/ actresses shows a multi-layered interplay of personal liking, religious sentiments, and cultural values. It shows all of these factors in continuous connection with personal opinion as well. It is easy to conclude that despite many problems facing the use of social media platforms, it can still be used to advance some positivity in society. It can be engaged in the celebration of shared cultural values and personal milestones. It can also bridge the difference between the life of celebrities and common people by providing the latter with the opportunity to have a peak into the former's lives. 

In conclusion, the sentiment analysis conducted on comments from marriage posts of Pakistani actors highlights the complex interplay between cultural context and public opinion. The predominantly positive sentiment observed in these comments demonstrates that, despite the challenges and potential pitfalls of online interactions, social media can serve as a platform for celebrating shared cultural values and personal milestones, such as weddings.

The presence of both positive and negative sentiments in the analyzed comments underscores the diversity of opinions within the online community. Positive comments emphasize a sense of shared joy and appreciation for cultural traditions, while negative comments reflect concerns about adherence to religious and cultural norms. This duality is indicative of the complexities that arise when public opinion is shaped by conservative societies where religion plays a significant role in defining social expectations.

Overall, the sentiment analysis of comments on marriage posts of Pakistani actors reveals that, while cultural and religious concerns may lead to negative sentiment in some cases, the majority of online users express support and appreciation for the shared cultural experiences of others. This finding highlights the potential for social media platforms to foster a sense of unity and celebration, even in the face of diverse opinions and perspectives.

References

• Agarwal, A., Xie, B., Vovsha, I., Rambow, O., & Passonneau, R. (2011). Sentiment Analysis of Twitter Data. Association for Computational Linguistic, 30–38. https://www.aclweb.org/anthology/W11-0705.pdf

• Balahur, A., Steinberger, R., Kabadjov, M., Zavarella, V., Van Der Goot, E., Halkia, M., ... & Pouliquen, B. (2011). Sentiment analysis in the news. arXiv preprint arXiv:1104.3874.
• Bakhshi, S., Shamma, D. A., & Gilbert, E. (2014). Faces engage us. https://doi.org/10.1145/2556288.2557403
•  Cambria, E., Schuller, B., Xia, Y., & Havasi, C. (2013). New avenues in opinion mining and sentiment analysis. IEEE Intelligent Systems, 28(2), 15–21. https://doi.org/10.1109/mis.2013.30
• De Choudhury, M., Sharma, S., & Kiciman, E. (2016). Characterizing Dietary Choices, Nutrition, and Language in Food Deserts via Social Media. In Proceedings of the 19th ACM Conference on Computer-Supported Cooperative Work & Social Computing. https://doi.org/10.1145/2818048.2819956
• Boudad, N., Faizi, R., Thami, R. O. H., & Chiheb, R. (2018). Sentiment analysis in Arabic: A review of the literature. Ain Shams Engineering Journal, 9(4), 2479–2490. https://doi.org/10.1016/j.asej.2017.04.007
• Hu, Y., Manikonda, L., & Kambhampati, S. (2014). What we Instagram: A first analysis of Instagram photo content and user types. Proceedings of the International AAAI Conference on Web and Social Media, 8(1), 595–598. https://doi.org/10.1609/icwsm.v8i1.14578
• Jansen, B. J., Zhang, M., Sobel, K., & Chowdury, A. (2009). Twitter power: Tweets as electronic word of mouth. Journal of the American Society for Information Science and Technology, 60(11), 2169–2188. https://doi.org/10.1002/asi.21149
• Kim, E., & Kim, H. Y. (2016). An analysis of user-generated contents about Korean Wave. Journal of the Korea Contents Association, 16(5), 354-364.
• Liu, B. (2012). Sentiment analysis and opinion mining. Synthesis Lectures on Human Language Technologies, 5(1), 1–167. https://doi.org/10.2200/s00416ed1v01y201204hlt016
• Loria, S., Keen, P., Honnibal, M., Yankovsky, R., Karesh, D., & Dempsey, E. (2014). TextBlob: Simplified text processing. Secondary TextBlob: Simplified Text Processing.
• Go, A., Bhayani, R., & Huang, L. (2009). Twitter sentiment classification using distant supervision. CS224N Project Report, Stanford University, 1(2009), 12.
• Mikolov, T., Sutskever, I., Chen, K., Corrado, G. S., & Dean, J. (2013). Distributed Representations of Words and Phrases and their Compositionality. Neural Information Processing Systems, 26, 3111–3119. http://ptrckprry.com/course/ssd/reading/Miko13.pdf
•  Pak, A., & Paroubek, P. (2010). Twitter as a corpus for sentiment analysis and opinion mining. Language Resources and Evaluation. https://www.researchgate.net/profile/Patrick_Paroubek2/publication/220746311_Twitter_as_a_Corpus_for_Sentiment_Analysis_and_Opinion_Mining/links/0deec52bb2a4cc37f8000000.pdf
•  Tumasjan, A., Sprenger, T., Sandner, P., & Welpe, I. (2010). Predicting Elections with Twitter: What 140 Characters Reveal about Political Sentiment. Proceedings of the International AAAI Conference on Web and Social Media, 4(1), 178–185. https://doi.org/10.1609/icwsm.v4i1.14009
•  Taboada, M., Brooke, J., Tofiloski, M., Voll, K., & Stede, M. (2011). Lexicon-Based Methods for Sentiment Analysis. Computational Linguistics, 37(2), 267–307. https://doi.org/10.1162/coli_a_00049