![](bg2.png)
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