Graphs and Key Players in an Educational Social Network

Fernando Calle-Alonso

, Vicente Botón-Fernández

, Dimas de la Fuente

, Carlos J. Pérez

Miguel A. Vega-Rodríguez

and Daniel de la Mata

Research & Development Department, ASPgems SL, Spain

Cátedra ASPgems, University of Extremadura, Spain

Department of Technologies of Computers & Communications, University of Extremadura, Spain

Department of Mathematics, University of Extremadura, Spain

Keywords: E-learning, Social Graph, Cooperative Learning, Key Player, Engagement, Dropout, NeuroK, Social Learning

Analytics.

Abstract: A new social learning graph tool has been proposed and implemented in NeuroK. This is an interactive

visualization of the relationships among students, based on their comments, favorites, mentions and rates in a

course or a learning unit. Some illustrative examples are provided showing its possibilities, challenges and

future potential of using the social graph tool. Social graph can help teachers and students to have an easy

visual image of their relationships, discovering the key players in the network and also the isolated students

in risk of dropping out. The teacher can use this information to reach the students “in risk” and try to retain

them with some motivating and engaging actions. The social graph and the indices that can be obtained from

it show up to be also a very good tool to analyse the development of the course and to help the teacher to

evaluate the students.

1 INTRODUCTION

E-learning is a current formative strategy that can

solve many problems from the traditional education

environment. But it can also have some regrets too

depending on the platform used, such as the isolation

of the students, the high dropout rate, or the

dependence on technology/internet. Depending on

the kind of learning process, the student can be more

or less involved and motivated, and the pros and cons

are different than with the traditional approach.

The learning process can be individual,

collaborative or cooperative. Online education

technology can support these three perspectives

(Schoor et al., 2015), but which kind of learning is

more effective? Technology may support both

individual learning and access to social networks to

allow the individual development, engagement, and

motivation within social relationships. McInnerney

and Roberts (2004) point out that there is a certain

overlap and inter-concept usage between cooperative

and collaborative learning and they have provided

definitions and a deep analysis. A way to distinguish

the three concepts is focusing on how the learning is

taking place. Individual learning is conducted alone,

collaborative learning depends on groups and

cooperative learning takes place in networks (Panitz,

1999).

Online cooperative learning (Paulsen, 2017) tries

to give the students a space for individual

development, and also the freedom to participate in

networks or communities to improve their knowledge

together.

This article presents a new tool showing a social

interactive graph, which is implemented in an

innovative e-learning platform called NeuroK (Calle-

Alonso et al., 2017). This platform is based on the

principles of neurodidactics, and on cooperative

learning. The relationships among students (and also

with the teacher) are shown with some illustrative

examples. A great number of visual interactive

features have been implemented and shown in this

paper, and the way to calculate the graph is explained.

This kind of social graph analysis is demonstrated to

be a good way to measure collaborative or

cooperative distance learning (Reffay and Chanier,

2003). Some e-learning platforms have tried to use

social networks to develop their learning process, but

most of them have taken advantage of the most

common applications, such as Facebook or Twitter,

Calle-Alonso, F., Botón-Fernández, V., de la Fuente, D., Pérez, C., Vega-Rodríguez, M. and de la Mata, D.

Graphs and Key Players in an Educational Social Network.

DOI: 10.5220/0006816605230527

In Proceedings of the 10th International Conference on Computer Supported Education (CSEDU 2018), pages 523-527

ISBN: 978-989-758-291-2

523

instead of developing and using their own. The

experiences in many cases have not been satisfactory,

and almost none of them have performed a social

analysis of the interactions among students.

In online education, it is very important to boost

the motivation and the engagement with the purpose

to reduce the dropouts (Willging and Johnson, 2009).

With the social graph it can be observed which

students are isolated or disconnected from the others,

and which are the key players (Zhou and Chen, 2015)

that act as influencers. The role of this influencers is

very important because they can spread information

in the network and reaching a lot of students. In this

paper an example of how to identify this kind of

student is also presented.

According to this information from the graph and

the role of the student in the network, the teacher

could grab one disconnected student and try to regain

his motivation. Also, the teacher could use this easy

to understand visual information as one of the factors

contributing to assess students, in this case according

to their participation in a course, or a lesson.

2 SOCIAL GRAPHS

Nowadays, the popularity of social networks has led

teachers to use them as a motivating factor (Rennie

and Morrison, 2013). Social learning networks

analytics are beginning to play an important role in e-

learning platforms (Buckingham Shum and Ferguson,

2012). The relationships among students of a social

network hide a lot of important information that can

be extracted with the right tools. This information can

be used to measure the participation and the cohesion

of a group of students. In addition, social networks

provide social learning graphs to capture and

represent the interactions and relationships that take

place among multiple entities in a learning

environment (Pardo, 2013).

Some teachers have tried to use some of the most

popular social networks (such as Facebook, Twitter

or LinkedIn) for educational purposes (Bosch, 2009).

However, most e-learning platforms do not have any

kind of social graph analysis or valuable measures to

properly evaluate the participants' relationships.

In this section we present the analysis of social

networks that is being applied to the cooperative e-

learning platform, NeuroK.

2.1 Implementation

Social graphs are considered as a global mapping of

the members within a social network and how they

relate to each other. In our case, the graph focuses on

an e-learning social environment. Nodes represent

students from a certain course or topic and the lines

linking the nodes represent a relationship between

them. In our social graph representation, the thicker

is the linking line between two nodes, the greater is

the number of communications between the two

students represented by the nodes. Due to the

different types of communications that may exist

between two students, each of them has been assigned

a different weight that affects the thickness of the line,

as we will see below.

The data obtained from NeuroK to build the social

graph and to calculate the social measures are mainly:

 Comments: these are the replies that students

make of a content published by another student,

creating a direct communication between them.

 Mentions: a student can be mentioned in

comments or tasks evaluations, creating a

social bond between the student mentioned and

the one who mentions.

 Favorites: sometimes a student can set as

favourite another student’s comment, pointing

out that they like its content.

 Rates: students can rate a content proposed by

another student as long as they justify their

assessment.

Each type of communication provides a different

score for the student who receives it. This score is the

one used as a weight when establishing the thickness

of the lines in the social graph, as it was mentioned

before.

For the representation of the graph, d3.js library

has been used. This library has some advantages such

as:

 It lets you make your visualizations in the way

you want (e.g. you can use avatars as node

representations).

 It allows you to add many other DOM

(Document Object Model) functions, like zoom

or pan functions for any graph you want.

 It supports a large amount of data.

 It has tools that make the connection between

data and graphics easy.

 It is as flexible as the client side web

technology stack.

 It allows you to build a graph based on social

gravity. This representation follows some rules

according to this gravity, such as the length of

a line or the attraction between nodes.

 It is easy to debug using the browser’s built-in

element inspector.

A2E 2018 - Special Session on Analytics in Educational Environments

524

The teacher and students can interact with the

social graph by doing any of the following actions

(see Fig. 1):

 Choosing the course or topic they want to be

represented in the graph.

 Filtering data by user: students and teacher/s or

only students.

 Filtering data by event type: comments,

mentions, favourites, rates or all events.

 Selecting the number of influencers: from 1 to

5 or no influencers.

 Displaying a tooltip when the mouse is over

any node.

 Moving the graph structure by dragging a node.

 Highlighting all the relationships of a node

when the mouse is over it.

 Displaying the neighbourhood of a node by

double clicking that node.

 Zooming in or out the entire social graph using

the mouse wheel.

Figure 1: Social graph filters.

The social graph represents the relative

importance of each student in NeuroK based on their

activity (comments, mentions, likes or ratings)

providing great information about the structure of the

network and how the students are related, as we will

see in the next subsection.

2.2 Study Case

Here a typical scenario that covers the concepts

mentioned above has been described in order to better

understand the system operation. We have tested

NeuroK through an existing topic called “Machine

Learning” to obtain a real dataset. There are five

students enrolled in the course and two teachers. In

order to build the social graph, all the information

from comments, mentions, favorites and rates has

been saved into the NeuroK database. That

information has been registered during a period of

one month. Once the dataset is ready, it is time to

navigate to the “Social Graph” view and set up the

filters. In this case, we establish the following

settings: Machine Learning topic, all users, all events

and no influencers.

After running our social graph approach over the

previous dataset, it generates the social graph that

appears in Fig. 2. It shows a global mapping of

students and teachers enrolled in the Machine

Learning topic and how they are related.

Figure 2: Social graph for Machine learning topic.

The nodes include the avatars of each student

profile (in this case we use the default avatar to

maintain their anonymity). They also have different

border colors depending on their role and how they

relate to each other. The students with social activity

have a green border. A student is considered to have

social activity when they send/receive comments,

mentions, favorites, and rates to/from other students.

The greater the social influence, the thicker and more

intense the green border. The metrics used to measure

this social influence will be discussed in the next

section. On the other hand, the isolated students with

no social activity have an orange border. Finally,

teacher nodes have no borders, since their activity

will not be evaluated.

The lines linking the nodes also highlight the

intensity of the relationship between two nodes. The

thickness of the line defines the intensity of the

relationship. With the proposed graph, a teacher can

have an overview of the participation in a course or

topic with a single chart. Some of the students may be

isolated, indicating that they are not participating in

the course or topic. The teacher should try to find out

why they are not using the platform. This is a simple

and effective way to identify students who may leave

the course before completing it.

Graphs and Key Players in an Educational Social Network

525

3 KEY PLAYERS IN

E-LEARNING

The information of the social network provided by the

graph would help in the identification of influencers

or key players, who may be good knowledge brokers.

Identification of a set of key players in a given social

network is of great interest in this context. Up to now,

most of the used algorithms for this task are based on

a single characteristic, however, in this case, it is

necessary to find a set of key players which can

perform well with respect to several characteristics of

interest. Within a social graph, students are

distributed into clusters. These clusters can be

analysed to find out what they are talking about and

what learning contents they are sharing and

commenting on. Clusters should not be isolated from

each other. The teacher's goal should be for all groups

to be connected, share information and collaborate

with each other. Some students will be related to

different groups. These students are often selected as

influencers or key players, as they can provide

information between groups and have different views

on the proposed materials and discussions. These key

players are very interesting to analyse, and teachers

could propose a specific line of action to help the rest

of the students thanks to their broad influence.

Different metrics have been used when assessing

the influence or social activity of a particular student:

betweenness centrality, global activity and Dijkstra’s

distance. The mean value of these three metrics

represents the influence of a node. With these three

metrics we manage to assess three fundamental

aspects of an influencer: direct communication with

the largest number of students (Dijsktra’s distance),

largest number of communications received or sent

(global activity) and communication between

different groups (betweenness centrality).

Let us go back to the study case from Subsection

2.2. Taking the same filter settings, but showing only

students and modifying the number of influencers to

1 as shown in Fig. 3, we can display the key player of

that topic.

4 CONCLUSIONS AND FUTURE

WORK

A new social graph tool has been proposed and

implemented in the e-learning platform NeuroK. It

has many visual customizations and possibilities to let

the teacher observe the participation of the students in

a course. With this tool the teacher is able to discover

the main relationships among the students and the

intensity of the relationships, which of them are key

players and who is disconnected and is a candidate to

drop out. Teachers can then try to re-engage isolated

students. Also they can use the social participation of

the students as part of their evaluation at the end of

the course.

In addition, groups of students can stand out showing

different behaviors and in the future, we classification

Figure 3: Social graph and key player for Machine learning

topic.

Fig. 4 shows the direct relationships of the

influencer by placing the mouse cursor over it.

Figure 4: Key player relationships.

A2E 2018 - Special Session on Analytics in Educational Environments

526

models to discover the main characteristics of each

group to deliver personalized recommendations.

Finally, the main indices obtained from the social

graph will be included in an evaluation book,

available for the teachers to help them to evaluate

students, and also, as a future work, an automatic

marking will be suggested.

ACKNOWLEDGEMENTS

This research has been supported by Ministerio de

Economía y Competitividad (Centro para el

Desarrollo Tecnológico Industrial, Contract IDI-

20161039), Junta de Extremadura (Contract AA-16-

0017-1, and projects GR15106 and GR15011),

Cátedra ASPgems, and European Union (European

Regional Development Funds).

REFERENCES

Bosch, T.E. 2009. Using online social networking for

teaching and learning: Facebook use at the University

of Cape Town. South African Journal for

Communication Theory and Research, 35(2), 185–200.

Buckingham Shum, S., & Ferguson, R. 2012. Social

Learning Analytics. Educational Technology &

Society, 15(3), 3–26.

Calle-Alonso, F., Cuenca-Guevara, A., de la Mata Lara, D.,

Sánchez-Gómez, J. M., Vega-Rodríguez, M. A., &

Sánchez, C. P. (2017). NeuroK: A Collaborative e-

Learning Platform based on Pedagogical Principles

from Neuroscience. In Proceedings of the 9th

International Conference on Computer Supported

Education (CSEDU 2017)-Volume Vol. 1, pp. 550-555.

McInnerney, J. M., & Roberts, T. S. 2004. Collaborative or

cooperative learning? Online collaborative learning:

Theory and practice, 203-214.

Panitz, T. 1999. Collaborative versus Cooperative

Learning: A Comparison of the Two Concepts Which

Will Help Us Understand the Underlying Nature of

Interactive Learning.

Pardo, A. 2013. Social learning graphs: combining social

network graphs and analytics to represent learning

experiences. International Journal of Social Media and

Interactive Learning Environments, 1(1), 43–58.

Paulsen, M. F. 2017. Cooperative online education.

In Seminar.net (Vol. 4, No. 2).

Reffay, C., & Chanier, T. (2003). How social network

analysis can help to measure cohesion in collaborative

distance-learning. In Designing for change in

networked learning environments (pp. 343-352).

Springer Netherlands.

Rennie, F., & Morrison, T. 2013. E-learning and social

networking handbook: Resources for higher education.

Routledge.

Schoor, C., Narciss, S., & Körndle, H. 2015. Regulation

during cooperative and collaborative learning: A

theory-based review of terms and

concepts. Educational Psychologist, 50(2), 97-119.

Willging, P. A., & Johnson, S. D. (2009). Factors that

influence students' decision to dropout of online

courses. Journal of Asynchronous Learning

Networks, 13(3), 115-127.

Zhou, J., & Chen, Y. J. (2015). Key leaders in social

networks. Journal of Economic Theory, 157, 212-235.

Graphs and Key Players in an Educational Social Network

527