In online social media platforms, users can express their ideas by posting original content or by adding comments and responses to existing posts, thus generating virtual discussions and conversations. Studying these conversations is essential for understanding the online communication behavior of users. This study proposes a novel approach to retrieve popular patterns on online conversations using network-based analysis. The analysis consists of two main stages: intent analysis and network generation. Users’ intention is detected using keyword-based categorization of posts and comments, integrated with classification through Naïve Bayes and Support Vector Machine algorithms for uncategorized comments. A continuous human-in-the-loop approach further improves the keyword-based classification. To build and understand communication patterns among the users, we build conversation graphs starting from the hierarchical structure of posts and comments, using a directed multigraph network. The experiments categorize 90% comments with 98% accuracy on a real social media dataset. The model then identifies relevant patterns in terms of shape and content; and finally determines the relevance and frequency of the patterns. Results show that the most popular online discussion patterns obtained from conversation graphs resemble real-life interactions and communication.

1. Marco Brambilla, Alireza Javadian Sabet,
and Amin Endah Sulistiawati
Conversation Graphs in
Online Social Media

2. Table of Content
 Introduction
 State-of-the-art in Conversation Graphs on Social Media
 Rsearch Questions and Contributions
 Case study
 Proposed methodology
 Intent Analysis
 Network Generation
 Pattern Identification
 Conclusion
 Refrences

3. Introduction
 The emergence of social media (SM) has profoundly changed the perspective
of communication, which resulted in a revolution in the way people interact
with each other [1].
 In online SM platforms, users can express their ideas by posting original
content or by adding comments and responses to existing posts, thus
generating virtual discussions and conversations.
 For reasons such as: interacting within the inner circle of friendship,
entertainment purposes, subscribing to news, knowledge sharing purpose on
online learning, and Q&A platforms [2],[3].
 Many companies adopt SM to utilize this growing trend to gain business
values [4].

4. State-of-the-art in Conversation Graphs on Social Media
 Ning et al. [5] utilize graph analysis to better support Q&A systems.
 Aumayr et al. [6] explore classification methods for recovering the reply
structures in forum threads.
 Cogan et al. [7] propose a method to reconstruct complete conversations
around initial tweets.
 Zayats and Ostendorf [8] predict the popularity of comments on Reddit
discussions.
 Kumar et al. [9] propose a mathematical model for the generation of basic
conversation structure to explore the model humans follow during online
conversations.
 Aragon et al. [10] investigate the impact of threading the messages instead of
linearly displaying them.

5. Research Questions
 RQ1: How to define appropriate graph models representing conversations on
SM?
 RQ2: How to reconstruct complex conversations structures when they are not
explicitly tracked by the social network platforms?
 RQ3: How to assign author intentions to posts and comments in the
conversations?
 RQ4: How to identify recurrent discussion patterns in conversation graphs?

6. Contributions
 A graph-based view on the discussions between social media contributors.
 Retrieve popular patterns on online conversations.
Intent Analysis
Network Generation
Pattern Identification

7. Case Study: YourExpo2015 Game Challenge*
 Long-running Live Event [11]
* http://www.socialmediaexpo2015.com/yourexpo/
 15,000 Photos  600,000 Actions
 100,000 Comments  80,000 Participants

8. Intent Analysis

9. Intent Analysis Pipeline

10. Bag of words with the most frequent occurrence words

11. Initial keywords for comment categories

12. Sentiment Analysis vs. Intent Analysis
 General sentiment analysis: Positive, Negative, and Neutral.
 Intent analysis: provides more understanding of the communication patterns
of the online users.

13. Network Generation

14. Network Generation (Design)
Graph visualization of a post on Social Media
Network design for Social Media platforms

15. Network Generation (Example)
Three posts and
respective relevant
nodes

16. Network Generation (Example)
A detailed conversation
with annotated intentions

17. Analysis Results

18. Statistical analysis of comments and conversations
The experiment is performed on 15,343 Instagram’s photos related to the case study

19. Frequency for each size of conversation
(number of comments)
frequency

20. Comment category distribution (min 30 comments)
Distribution of comment categories on conversations having minimum 30 comments
(number of comments)

21. Comment category distribution (7-29 comments)
Distribution of comment categories on conversations having number of comments between 7 and 29
(number of comments)

22. Temporal Patterns
3D representation of the conversation size, period, and frequency
(number of comments)

23. Content Relationships (2-Nodes pattern)
The frequency of the comment-reply relationship for categories

24. Content Relationships (3,4-Nodes pattern)
Expanding from the
2-node patterns with
more than 1,000
occurrences.
thank → positive
positive → positive
thank → thank
thank → food
thank → greeting

25. Reply time in thank  thank  positive pattern

26. Reply time in thank  thank  thank  positive pattern

27. Number of user in top conversation patterns
The number of users that join the top conversation patterns
3 Nodes 4 Nodes

28. Conclusion
 By identifying the relationships among all comments on an SM post:
 Retrieve the discussions
 Construct the conversation graphs
 By mining the constructed conversation graphs:
 Identify the popular conversation patterns
 Provide interesting insights into the SM users’ preferences and behaviors
 Useful for designing conversational agents

29. References
[1] Qualman, Erik. Socialnomics: How social media transforms the way we live and do business. John Wiley & Sons, 2012.
[2] Al-Atabi, Mushtak, and Jennifer DeBoer. "Teaching entrepreneurship using massive open online course (MOOC)." Technovation 34.4
(2014): 261-264.
[3] Vasilescu, Bogdan, et al. "How social Q&A sites are changing knowledge sharing in open source software communities." Proceedings
of the 17th ACM conference on Computer supported cooperative work & social computing. 2014.
[4] Dong, John Qi, and Weifang Wu. "Business value of social media technologies: Evidence from online user innovation communities."
The Journal of Strategic Information Systems 24.2 (2015): 113-127.
[5] Yang, Jaewon, Julian McAuley, and Jure Leskovec. "Community detection in networks with node attributes." 2013 IEEE 13th
international conference on data mining. IEEE, 2013.
[6] Aumayr, Erik, Jeffrey Chan, and Conor Hayes. "Reconstruction of threaded conversations in online discussion forums." Proceedings of
the International AAAI Conference on Web and Social Media. Vol. 5. No. 1. 2011.
[7] Cogan, Peter, et al. "Reconstruction and analysis of twitter conversation graphs." Proceedings of the First ACM International Workshop
on Hot Topics on Interdisciplinary Social Networks Research. 2012.
[8] Zayats, Victoria, and Mari Ostendorf. "Conversation modeling on reddit using a graph-structured lstm." Transactions of the Association
for Computational Linguistics 6 (2018): 121-132.
[9] Kumar, Ravi, Mohammad Mahdian, and Mary McGlohon. "Dynamics of conversations." Proceedings of the 16th ACM SIGKDD
international conference on Knowledge discovery and data mining. 2010.
[10] Aragón, Pablo, Vicenç Gómez, and Andreaks Kaltenbrunner. "To thread or not to thread: The impact of conversation threading on
online discussion." Proceedings of the International AAAI Conference on Web and Social Media. Vol. 11. No. 1. 2017.
[11] Brambilla, Marco, Alireza Javadian Sabet, and Marjan Hosseini. "The role of social media in long-running live events: The case of the
Big Four fashion weeks dataset." Data in Brief 35 (2021): 106840.

30. THANKS!
QUESTIONS?
Marco.Brambilla@polimi.it Alireza.Javadian@polimi.it Amin.Endahsulistiawati@mail.polimi.it
@MarcoBrambi @ArjSabet
Brambilla Marco, Javadian Alireza, Sulistiawati Amin (2021) Conversation Graphs in Online Social Media.
In: Int. Conf. on Web Engineering, ICWE 2021. LNCS, vol. 12706, pages 97-112. Springer, Cham.
doi: 10.1007/978-3-030-74296-6_8
http://datascience.deib.polimi.it/