Available online at www.sciencedirect.

com

ScienceDirect
Procedia - Social and Behavioral Sciences 237 (2017) 792 – 797

7th International Conference on Intercultural Education “Education, Health and ICT for a
Transcultural World”, EDUHEM 2016, 15-17 June 2016, Almeria, Spain

A comparison of intercultural student communities in Online Social

Networks
Carlo Giglio* & Roberto Palmieri
a
Department of Mechanical, Energy and Management Engineering, University of Calabria, Building 42/C, Rende (CS), 87036, Italy

Abstract

This work is geared to analyze informal learning processes in student conferences. In particular, it compares social network
interactions occurring in conference-related Facebook pages - i. e. Taiwan-America Student Conference (TASC), Japan-America
Student Conference (JASC) and Korea-America Student Conference (KASC) pages - within a period starting 30 days before and
finishing 30 days after the application deadlines. This empirical study has been realized by adopting open source visualization tools
and techniques freely available on the software market in order to perform Social Network Analysis (SNA) in a transparent and
reproducible way. Such an analysis provides interesting information on interaction dynamics, emerging hot topics and sub-group
formation of attending students.
© 2017
© 2016TheTheAuthors.
Authors.Published
Publishedbyby Elsevier
Elsevier Ltd.Ltd.
This is an open access article under the CC BY-NC-ND license
Peer-review under responsibility of the organizing committee of EDUHEM 2016.
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of EDUHEM 2016.
Keywords: Social Network Analysis; Informal Learning; Online Communities; Learning Patterns; Facebook;

Theme field: ICTs in learning as developmental processes

1. Introduction

Creating and enhancing personal and professional relationships among young people – e. g. students attending
international conferences – leads to better results in terms of knowledge generation, updating and sharing (Palmieri &
Giglio, 2015a; Palmieri & Giglio, 2015b; Palmieri & Giglio, 2015c). This is even more true when it comes to foster
knowledge exchange among new generations coming from different countries. As a matter of fact, such conferences

* Carlo Giglio. Tel.: +39-0984-494775; fax: +39-0984-494775.

E-mail address: carlo.giglio@unical.it

1877-0428 © 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of EDUHEM 2016.
doi:10.1016/j.sbspro.2017.02.148
 Carlo Giglio and Roberto Palmieri / Procedia - Social and Behavioral Sciences 237 (2017) 792 – 797 793

are geared to encourage attendees to pursue a shared vision of a peaceful future and to develop informal collaborations
agreement in order to share ideas and concepts. Therefore, exchanging knowledge fosters most of innovations aimed
at creating a better world (Yusuf, 2009; de Castro, Rodrigues, Esteves, & da Rosa Pires, 2000; Burton-Jones, 2001;
Iammarino, 2005; Palmieri & Giglio, 2014). Such events provide students also with the opportunity to improve their
skills or develop new ones. This way, attendees may obtain competitive advantages in the job market and, hence,
higher performance in both social and professional activities (Takeuchi, 2006; Di Pietro & Anoruo, 2006).
However, analyzing key issues related to social media interactions and their effects on informal learning of
conference attendees may be difficult. The main problem in this field is associated with the choice of the most adequate
visualization tools and techniques due to the abundance of software solutions available on the market (Khan & Khan,
2011). Such an issue emerges also from the even bigger volume of data and information available in different contexts
(North, 2012). Hence, it affects the data elaboration and interpretation process and, thus, influences the proper
understanding of the case study at hand. As a matter of fact, visualization-related topics are worth attention of both
academic and professional study groups (Kban & Khan, 2011). Each case study can lead to partially different
conclusions based on the perspective of analysis adopted: different visualization approaches may highlight different
factors, thus, affecting the overall understanding of analysts and scientists (Ware, 2004; Spence, 2001). Currently, the
design of the visualization process, the visualization methods clustering process, and possible combinations of
different methods represent the most relevant issues in this field. Therefore, it is important to discuss about theoretical
issues and definitions of visualization methods and processes.
Visualization is defined as the representation of phenomena in a similar way to the original perception through
proper computer-based tools and techniques (Khan & Khan, 2011; Card, Mackinlay, & Shneiderman, 1999), which
fosters knowledge acquisition processes (Teyseyre & Campo, 2009; Khan &Khan, 2011). In fact, a richer and more
intuitive representation can be obtained by means of graphical description tools (Tufte, 1997), which facilitate ranking
and clustering of available data from different perspectives in order to perform the required analysis (Kowalski &
Maybury, 2002; Butler, Almond, Bergeron, Brodlie, & Haber, 1993). The design of the visualization process is
composed of a six-step model (Chittaro, 2006):

x Mapping refers to the processes of encoding and rendering in visual form;

x Selection is related to the separation between useful and useless data, which depends on the specific goals of the
analysis at hand and may affect its success;
x Presentation refers to the way data and information are visually conveyed to all those potentially interested in it;
x Interactivity concerns the possibility for users to work on data and information through available tools and
techniques;
x Accessibility and usability emphasize the “human factor” as they aim at satisfying special needs;
x Evaluation concerns the comparison between expected and actual results by using questionnaires, interviews, focus
groups, controlled experiments, cognitive walkthroughs and expert reviews (Teyseyre and Campo, 2009).

The main data visualization methods are briefly detailed (Khan & Khan, 2011):

x Table is an intuitive and flexible format combining variables, factors and values;
x Pie chart visualizes data by means of slices associated with variables, factors and (percentage) values;
x Bar chart is a flexible tool for data visualization;
x Histogram clusters variables and factors into aggregated categories;
x Line chart visualizes data as a sequence of points associated with the corresponding values;
x Area chart reports available data in a bounded area;
x Scatter plot represents points in Cartesian coordinates;
x Bubble chart is a variant of scatter plot whereby bubble points are also endowed with the value of the diameter;
x Multiple Data Series is a mix of the aforementioned methods.

The objectives of this study are detailed in Section II. The choice of methods, tools and techniques adopted in this
work is detailed and justified in Section III. Section IV introduces and discuss empirical results about the data
794 Carlo Giglio and Roberto Palmieri / Procedia - Social and Behavioral Sciences 237 (2017) 792 – 797

elaboration of social media activities occurring in the three Facebook pages at hand. Finally, Section V ends with
conclusions.

2. Objectives

Authors aim at analyzing informal learning dynamics by means of visualization techniques in order to compare
interactions happening within three pre-event conference-related Facebook pages. Moreover, this work is geared to
identify in advance emerging conference hot topics in order to nurture a favorable context during the event. Finally,
such studies contribute to track sub-networks formation processes before the starting of the conference. This way,
organizers are endowed with more detailed information, which may help setting up proper scenarios fostering the
clustering process of attendees. In turn, organizers may ensure that such clusters of participant experience higher level
of satisfaction since each sub-group is characterized by a number of interests in common.

3. Methodology

Conferences foster knowledge generation, updating and sharing activities, which are generally planned by
organizers and arise from cognitive conflicts (Engeström, 2000). In addition, in international student conference
settings, participants bring their own cultures and, hence, they combine them. This way, attendees work in a
transcultural context characterized by significant Asian-American cultural exchanges. Such exchanges may lead also
to possible issues - e. g. collaboration, development of a shared vision, communication difficulties -, which may be
(partially) solved through ad hoc activities triggered by event organizers (Aramo-Immonen, Jussila, & Huhtamäki,
2014). Therefore, conference rules and event settings influence attendee behavior. Also the adoption of emerging
technologies and tools may enhance people engagement. In light of this, online communities are considered as the
most reliable and effective settings concerning the possibility for participants to clearly express and exchange ideas
and thoughts (Aramo-Immonen, Jussila, & Huhtamäki, 2014). It is even more true when ad hoc efforts - which are
devoted to promote collaboration and cultural exchanges in advance - are related to student communities characterized
by significant intercultural differences (Jussila, Huhtamäki, Kärkkäinen, & Still, 2013). Finally, online sources
associated with conference-related communities generally provide the most reliable data in order to perform this kind
of study (Card, Mackinlay, & Shneiderman, 1999; Benbasat, Goldstein, & Mead, 1987).
From a methodological perspective, three conference-related communities on Facebook are taken into account
through data visualization and analysis (Ware, 2004) tools. Namely, the Taiwan-America Student Conference
(hereinafter TASC), the Japan-America Student Conference (hereinafter JASC) and the Korea-America Student
Conference (hereinafter KASC) are considered. All of them will be held in July-August 2016. The data gathering
process covers Facebook activities over a temporal horizon starting 30 days before and finishing 30 days after the
application deadlines. The proposed methodological framework (Palmieri & Giglio, 2015a; Palmieri & Giglio, 2015b;
Palmieri & Giglio, 2015c) is compliant with the data science research approach (Hey, Tansley, & Tolle, 2009) and the
research methods applied to online sources (Davenport, 2014). The communities taken into account for TASC, JASC
and KASC count on 3,306 likers, 1,960 users and 745 users, respectively.
The data gathering process relies on the “page data”, “page like network” and “search” modules provided by the
NetVizz app v1.3 (Rieder, 2013), which has been created by Professor Bernhard Rieder for research purposes. It
represents a reliable system in order to collect data from online sources and elaborate them through other data
visualization software solutions (Rieder, 2013) like Gephi (Bastian, Heymann, & Jacomy, 2009), which ensures also
the possibility to adopt usable and scalable solutions, ad hoc algorithms, options and features. Moreover, both NetVizz
and Gephi are open source solutions and ensure that research results presented in this study are transparent and
reproducible.

4. Results

Gephi provides an information about the graph types, which are directed in the three case studies at hand. Moreover,
it identifies 278 nodes and 710 directed edges for the TASC graph, 191 nodes and 353 directed edges for the JASC
 Carlo Giglio and Roberto Palmieri / Procedia - Social and Behavioral Sciences 237 (2017) 792 – 797 795

one, and 64 nodes and 138 directed edges for the KASC one. Afterwards, graph density is analyzed. In fact, it helps
detecting graph completeness (Palmieri & Giglio, 2015a; Palmieri & Giglio, 2015b; Palmieri & Giglio, 2015c) and,
together with strongly and weakly connected components - hereinafter SCC and WCC, respectively - (Tarjan, 1972),
it contributes to identify possible sub-groups of attendees. Modularity measures are calculated by means of adopting
the same standard parameters and resolution values (Blondel, Guillaume, Lambiotte, & Lefebvre, 2008; Lambiotte,
Delvenne, & Barahona, 2009) in order to prove further the existence of possible sub-networks detections. Finally,
shortest paths, network diameters, average path lengths, HITS, centrality (Brandes, 2001; Kleinberg, 1999) and
PageRank (Brin & Page, 1998) measures are determined since they may be helpful in order to deepen general graph
properties like those related to possible sub-networks and connected components. The main data collected and
extracted and the corresponding measures are synthetically reported in Table 1.

Table 1. Data visualization from Gephi.

Data/Conference KASC JASC TASC
Number of nodes 64 191 278
Number of directed edges 138 353 710
Graph density 0.034 0.010 0.009
Weakly Connected Components (WCC) 1 7 2
Strongly Connected Components (SCC) 64 191 278
Modularity 0.396 0.456 0.375
Modularity with resolution 0.396 0.456 0.375
Number of communities 4 15 13
Network diameter (ND) 1 1 1
Average path length (APL) 1.0 1.0 1.0
Betweenness Centrality Distribution (BCD) [60; 70] [185; 200] [275; 280]
count range for values tending to 0
Betweenness Centrality Distribution (BCD) [0; 0+ε] [0; 0+ε] [0; 0+ε]
count range for values tending to 1
Closeness Centrality Distribution (CCD) count [15; 20] [40; 45] [50-ε; 50]
range for values tending to 0
Closeness Centrality Distribution (CCD) count [45-ε; 45] [145; 150-ε] [225; 235]
range for values tending to 1
Eccentricity Distribution (ED) count range for [15; 20] [40; 45] [50-ε; 50]
values tending to 0
Eccentricity Distribution (ED) count range for [45-ε; 45] [145; 150-ε] [225; 235]
values tending to 1
PageRank Distribution count range for values [0+ε; 45-ε] [0; 155] [0; 235]
tending to 0
PageRank Distribution count range for values [0; 0+ε] [0; 0+ε] [0; 0+ε]
tending to 1

First, the comparison of current data with previous measures related to the same conferences held in 2015 depicts
an increase of users in all of the case studies at hand (Palmieri & Giglio, 2015a; Palmieri & Giglio, 2015b; Palmieri
& Giglio, 2015c). However, the highest density level is detected in KASC, which has also the most limited number of
users and likers among the three conference-related pages at hand. Moreover, absolute values of density do not provide
a positive assessment of the triggering activities planned by the organizers. Despite such a negative evaluation, there
is a constant trend in terms of ability of organizers to involve newcomers. In fact, graph density values are very close
or slightly higher if compared to those calculated in 2015. Therefore, newcomers liking the conference pages and
interacting with them are characterized by the same (or slightly improved) engagement rates revealed in 2015.
796 Carlo Giglio and Roberto Palmieri / Procedia - Social and Behavioral Sciences 237 (2017) 792 – 797

Afterwards, strongly and weakly connected components (hereinafter SCC and WCC, respectively) are considered
(Tarjan, 1972) in order to understand whether pre-event conference sub-networks exist and to what degree they act as
distinct communities. Moreover, it is possible to analyze intra-component bounds in order to deepen what the
conditions are for interactions within existing sub-networks. The number of WCC is lower than in 2015 for KASC, in
compliance with the slightly increased graph density, but still absolute values are not sufficient in terms of attributing
a higher effectiveness to triggered activities planned by organizers. This means that every vertex in KASC graph is
more reachable from every other vertex than in the past. This means also that participants share the same interests and
there are no longer sub-groups with different perspectives, thoughts and ideas. It is proved also by a very small increase
in terms of number of communities, which is compliant with a reduction of modularity values. Changes in centrality
and PageRank distributions are coherent. On the contrary, JASC reveals an increase in terms of overall WCC and an
even more significant increase in terms of communities, which is compliant with centrality and PageRank distributions,
and with a small reduction of modularity values. In this case, attendee engagement is driven by ad hoc triggering
activities in the social media environment, which are devoted to specific sub-groups of participants. Therefore,
organizers should focus on finding and sharing contents on target hot topics for each sub-network. Finally, TASC
conference proves to be more stable than others in terms of number of communities detected and WCC if compared
to the simultaneous increase of online users. Such a constant trend is compliant with changes in centrality and
PageRank distributions, slightly increased modularity values, and a slightly decreased graph density measure.

5. Conclusions

The comparison of the conference-related pages at hand helps deepening attendee interactions, identifying pre-
event sub-groups formation, and determining in advance emerging hot topics. Ultimately, it contributes to the
development of human behaviour in this context. Moreover, it provides event organizers with a useful approach in
order to foster a breeding environment devoted to the development of emerging hot topics also during the conference.
In fact, event organizers could exploit the useful data provided by such an analysis in order to share possible ad hoc
contents with the target sub-communities both in the pre-event phase and during the conference. Therefore, this may
lead to a higher user satisfaction and, ultimately, to a higher event success rate, thus, achieving organizers goals.
Despite some differences about the increasing or decreasing trends of the conferences at hand, altogether, the case
study shows how the intercultural communities of students attending the international conferences at hand are not
characterized by an overall distinction among possible connected components and detected communities. In light of
this, the three Asian-American conferences prove how intercultural exchanges and attendee interactions are not
affected by a strong separation driven by some shared cultural, research or learning interests. On the contrary, this
study proves how intercultural Asian-American student conferences tend to ensure a significant level of user
engagement, which makes it possible for each online community to interact and keeps on staying close to each other
attendees, thus, gathering participants around the general topic of the event and avoiding most of inclinations towards
separation.
Data gathered from NetVizz and elaborated through Gephi reveal a lacking of potentially useful data in order to
identify emerging hot topics. This issues comes from very recent changes in privacy settings and policy adopted by
Facebook. Current privacy settings - e. g. user data anonymization - make it impossible also to determine whether
there is an overlapping between online sub-communities and real-world sub-groups. Despite such a policy makes it
very difficult to retrieve data for research purposes, this study shows how pre-event activities planned by conference
organizers still can be analyzed by means of the proposed research design. Therefore, it allows to deepen informal
learning process in intercultural student conferences and to track the emergence of possible sub-networks gathered
around specific hot topics and shared research interests. Ultimately, the proposed framework of analysis ensures the
possibility to overcome the aforementioned exogenous factors like changes in privacy policy operated by Facebook.
Hence, this work relies on a valuable research methodology, which is not influenced by such exogenous changes.
From the methodological point of view, the theoretical framework for data analysis relies on the adoption of new
and advanced data visualization tools and techniques. Moreover, the new approach is applied to a specific field that is
the comparison of informal learning dynamics in online social network communities of three Asian-American student
conferences. This way, the proposed study provides also a guidance in order to analyze case studies related to the
 Carlo Giglio and Roberto Palmieri / Procedia - Social and Behavioral Sciences 237 (2017) 792 – 797 797

research field at hand. Finally, this work may encourage field scholars to create more advanced sub-communities
detection methods and tools, together with conference hot topic prediction systems in the pre-event phase.

Bibliography

Aramo-Immonen, H., Jussila J., & Huhtamäki, J. (2014). Visualizing Informal Learning Behavior from Conference Participants Twitter Data, 2nd
International Conference on Technological Ecosystems for Enhancing Multiculturality – TEEM ‘14, pp. 603-610.
Bastian, M., Heymann, S., & Jacomy, M. (2009). Gephi: An Open Source Software for Exploring and Manipulating Networks, 3rd International
ICWSM Conference.
Benbasat, I., Goldstein, D. K., & Mead, M. (1987). The case research strategy in studies of information systems. MIS Quarterly, 369–386.
Blondel, V. D., Guillaume, J.-L., Lambiotte, R., & Lefebvre, E. (2008). Fast unfolding of communities in large networks. Journal of Statistical
Mechanics: Theory and Experiment, 10, 1000.
Brandes, U. (2001). A Faster Algorithm for Betweenness Centrality. Journal of Mathematical Sociology 25:2, 163-177.
Brin, S., & Page, L. (1998). The Anatomy of a Large-Scale Hypertextual Web Search Engine, 7th International Conference on the World Wide
Web (WWW1998), pp. 107-117.
Burton-Jones, A. (2001). The knowledge supply model: a framework for developing education and training in the new economy. Education and
Training.
Butler, D. M., Almond, J. C. Bergeron, R. D., Brodlie, K. W., & Haber, A. B. (1993). Visualization reference models, 4th IEEE Conference
Visualization, pp. 337-342.
Card, S. K., Mackinlay, J. D., & Shneiderman, B. (1999). Readings in information visualization: using vision to think. Morgan Kaufmann Pub.
Chittaro, L. (2006). Visualizing information on mobile devices. ACM Computer, 39:3, 40-45.
Davenport, T. (2014). Big Data at Work: Dispelling the Myths, Uncovering the Opportunities. Harvard Business Review Press.
de Castro, E. A., Rodrigues, C., Esteves, C., & da Rosa Pires, A. (2000). The triple helix model as a motor for the creative use of telematics.
Research Policy. Elsevier.
Di Pietro, W., & Anoruo, E. (2006). Creativity, innovation, and export performance. Journal of Policy Modeling. Elsevier.
Engeström, Y. (2000). Activity theory as a framework for analyzing and redesigning work. Ergonomics, 43:7, 960–974.
Hey, A. J., Tansley, S., & Tolle, K. M. (2009). The fourth paradigm: data-intensive scientific discovery. Microsoft Research.
Iammarino, S. (2005). An evolutionary integrated view of Regional Systems of Innovation: concepts, measures and historical perspectives.
European planning studies.
Jussila, J., Huhtamäki, J., Kärkkäinen, H., & Still, K. (2013). Information visualization of Twitter data for co-organizing conferences, 17th
International Academic MindTrek Conference: Making Sense of Converging Media. Tampere, Fnland.
Khan M., & Khan, S. S. (2011). Data and information visualization methods, and interactive mechanisms: a survey. International Journal of
Computer Applications, 34:1.
Kleinberg, J. M. (1999). Authoritative Sources in a Hyperlinked Environment. Journal of the ACM 46:5, 604–632.
Kowalski, G. J. & Maybury, M. A. (2002). Information storage and retrieval system, theory and implementation. Springer.
Lambiotte, R., Delvenne, J.-C-, & Barahona, M. (2009). Laplacian Dynamics and Multiscale Modular Structure in Networks. IEEE Transactions
on Network Science and Engineering, 1:2, 76-90.
North, C. (2012). Information Visualization. In G. Salvendy (Ed.), Handbook of Human Factors and Ergonomics (4th ed.). New York: John Wiley
& Sons (pp. 1209-1236).
Palmieri, R., & Giglio, C. (2014). Seeking the stakeholder-oriented value of innovation: a CKI perspective. Measuring Business Excellence, 18:1,
35-4.
Palmieri, R., & Giglio, C. (2015a). Informal Learning in Online Social Network Environments: An Evidence from an Academic Community on
Facebook. The Turkish Online Journal of Educational Technology.
Palmieri, R., & Giglio, C. (2015b). Using Social Network Analysis for a Comparison of Informal Learning in Three Asian-American Student
Conferences. The Turkish Online Journal of Educational Technology.
Palmieri, R., & Giglio, C. (2015c). Visualizing Social Network Data: a comparative study of Asian-American student conferences, 2nd International
Workshop on Social Networks Analysis, Management and Security - SNAMS 2015 - in conjunction with the 3rd IEEE International Conference
on Future Internet of Things and Cloud - FiCloud 2015. Rome, Italy.
Rieder B. (2013). Studying Facebook via data extraction: the Netvizz application, 5th Annual ACM Web Science Conference - WebSci '13 (pp.
346-355). New York: ACM.
Spence, R. (2001). Information visualization. AddisonWesley.
Takeuchi, H. (2006). The new dynamism of the knowledge creating company. In H. Takeuchi & T. Shibata (Eds.), Japan, moving toward a more
advanced knowledge economy: advanced knowledge-creating companies (Volume 2). Washington: World Bank. 1-9.
Tarjan, R. (1972). Depth-First Search and Linear Graph Algorithms. SIAM Journal on Computing, 1:2, 146–160.
Teyseyre, A. R., & Campo, M. R. (2009). An overview of 3D software visualization. IEEE Transactions on Visualization and Computer Graphics,
15:1.
Tufte, E. R. (1997). Visual explanations: images and quantities, evidence and narrative. Graphics Press.
Ware, C. (2004). Information Visualization: Perception for Design. Morgan Kaufmann, Elsevier.
Yusuf, S. (2009). From creativity to innovation. Technology in Society. Elsevier.