Dr. Anatoliy Gruzd
Canada Research Chair &
Director of Research
OurTeam
Philip Mai, JD
Director of Business &
Communications
6-10 Undergraduate,
Master /MBA & PhD
students (Business, CS,
Sociology, Information &
Psychology)
Drs. Jenna Jacobson &
Priya Kumar
Post Doctoral Fellows
Collaborators from across
Canada, the US, UK,
Netherlands, HK, Korea,
and Brazil.
2
About the
Social Media
Lab
SocialMediaLab.ca
@SMLabTO

Conference & Workshops
400+ Authors * 250+ Attendees * 28 Countries
Annual
International
Conference on
Social Media &
Society
#SMSociety
SocialMediaAndSociety.org

Download Slides and Practice Dataset
http://bit.ly/asac18
Anatoliy Gruzd (@gruzd) SOCIAL NETWORK ANALYSIS 4

Outline
◦ Brief Introduction to SNA
◦ Case Study: Organizational Networks
Hands-on part with Gephi:
◦ Sample Dataset: Massively Open Online Course for
Educators (MOOC-Eds)
◦ Exploratory Analysis using Network Visualization

Gephi
Network visualization, data preparation, exploration
Download from
https://gephi.org/users/download/
Requires Java JDK
http://www.oracle.com/technetwork/java/javase/downloads/jdk8-downloads-2133151.html

SNA Text
http://faculty.ucr.edu/~hanneman/nettext/
7

8
Social Network Analysis (SNA)
Nick
Rick
Dick
• Nodes = People
• Ties = Edges = Relations
Anatoliy Gruzd
(@gruzd)
Social Network Analysis

9
With networks we can answer questions
such as …
• What group/individuals stand out?
• Are there important connections?
• What is the “health” of the organization?
• How different are two groups/individuals or the same
group at two different times?
Anatoliy Gruzd
(@gruzd)

10
Using SNA in the learning context
• Identify students who might need extra attention/help from the
instructor
• Find active group members who often take a leadership role in a
group
• Identify peer-help
Student Instructor
Student
Group
LeaderStudent
Student

12
Directed vs Undirected Networks
Example: Communication Network (directed)
Nick
Rick
Dick
• Nodes = People
• Ties = “Who talks to whom”
• Tie strength (weight)= The
number of messages
exchanged between individuals
mutual/
reciprocal
Anatoliy Gruzd
(@gruzd)

Case Study
Organizational Networks

N is for Network: Mapping
Organizational Changes
Nancy Steffen-Fluhr, Regina Collins, Babajide Osatuyi,
Anatoliy Gruzd
NJIT-led and NSF-funded project (2009-2013)
Advancing Women at
NJIT through
Collaborative Research
Networks
14

“Universities and corporations are not merely buildings
and balance sheets…. They are relational entities--webs
of interaction and perception whose complex structure is
largely invisible to the people embedded in them.” (O'Reilly, 1991)
Anatoliy Gruzd
(@gruzd)
Social Network Analysis 15

Network structure drives institutional change… facilitating
(or retarding) innovation—maintaining (or altering) norms,
including norms of gender and race.
2000 2005 2008
Anatoliy Gruzd
(@gruzd)

“Network inequality creates and reinforces inequality of opportunity”
(Christakis & Fowler, 2009)
Understanding network dynamics is especially important for underrepresented
minorities and women who can easily spend their entire careers on the periphery,
far away from the flow of information at the core.
17

Being IN the Loop means:
• access to unpublished research
• invitations to join grant initiatives
• the opportunity to vet one’s work
• support for intellectual risk-taking
Being OUT of the loop
makes it harder to
accumulate social capital
which, in turn, has a
negative effect on
retention and
advancement. 18

NETWORK CENTRALITY
“The more paths that connect you to other people in your network, the more
susceptible you are to what flows within it.” (Christakis & Fowler, 2009)
Hypothesis: If women faculty members are less centrally
located than male faculty, they will incur greater information-
foraging costs and have fewer opportunities to signal their
value as organizational players, a difference that may
constitute a structural constraint for advancement.
Anatoliy Gruzd
(@gruzd)

NJIT ADVANCE is using SOCIAL NETWORK ANALYSIS
(SNA) to create new mapping tools that will give junior
faculty & their mentors an aerial view of the organizational
landscape …a “GPS System for Career Management.”
“Can’t see the forest
for the trees.” SNA Mapping
Anatoliy Gruzd
(@gruzd)

Phase One:
Building a Faculty Publications Database (DB)
2208 author names 7225 publications
Primary data acquisition method: data-mining of Scopus
DB Interface Modules:
Author/Publications Co-Authors Statistics
Anatoliy Gruzd
(@gruzd)

Phase Two:
Preparing to Map Changing Network Connections Among
NJIT Co-Co-authors
2000-2008
463 tenured/tenured-track STEM faculty
Mapping Tools:
NJIT DB UCINET ORA
• UCINET https://sites.google.com/site/ucinetsoftware/downloads
• ORA (Organizational Risk Analyzer) from CMU
http://www.casos.cs.cmu.edu/projects/ora/

Phase Three:
Using UCINET to test hypotheses about network structure,
collaboration, and career advancement.
Defining SNA Terms:
DEGREE CENTRALITY indicates well-connected people who can directly
reach many people in the network.
BETWEENNESS CENTRALITY reflects the extent to which an individual
has the ability to control the flow of information in the network.
EIGENVECTOR CENTRALITY reflects the extent to which an individual is
connected to well-connected people in the network.
Anatoliy Gruzd
(@gruzd)

Homophily
Male faculty are much more likely to co-author with other male
faculty than with women faculty
Anatoliy Gruzd
(@gruzd)

Network Centrality and Retention
For women faculty, Eigenvector centrality was a much stronger predictor of
retention than number of publications
Female Eigenvector Centrality = Retention
Anatoliy Gruzd
(@gruzd)

Network Centrality and Retention
For women faculty, Eigenvector centrality was a much stronger predictor of
retention than number of publications
Female Eigenvector Centrality = Retention
Collaboration and Advancement in Rank
Assistant and associate professors who co-authored more with other NJIT
faculty members exhibited greater upward movement in rank than those who
co-authored less.
More collaboration = Rise in academic rank (promotion)
Anatoliy Gruzd
(@gruzd)

Phase Four: Data Visualization
Using ORA to Create and Analyze Network Maps
Defining Terms:
Ego Network: a focal node and the nodes to whom it is directly connected (alters)
plus the ties among the alters.
Anatoliy Gruzd
(@gruzd)

Phase Four: Data Visualization
Using ORA to Create and Analyze Network Maps
Defining Terms:
Ego Network: a focal node and the nodes to whom it is directly connected (alters)
plus the ties among the alters.
Whole-Network Analysis maps “the occurrence and non-occurrence of relations
among all members of a population” (Garton,1997)
Anatoliy Gruzd
(@gruzd)

SNA Can Support Institutional Transformation
Bibliometric data —more and more easily accessible on a
national/global scale—is a valid proxy for real-world faculty
networks.
Drawing on such data, university policy makers can use new
SNA tools to…
▪ track changes in organizational health,
▪ identify emerging leaders or isolated backwaters
▪ compare the relative advancement of selected
groups/individuals.
Anatoliy Gruzd
(@gruzd)

Hands-On Part: Gephi

Massive Open Online Courses (MOOCs)
▪ A large scale reimagining
of traditional online
courses
▪ A typical MOOC consists
of 1,000+ students
▪ Since “The Year of the
MOOC” (NY Times, 2012),
interest has been steadily
rising

SNA may help to …

Practice Dataset: MOOC-Eds
‘MOOC-Eds are designed
specifically for professional
educators and follow the
guidelines for effective
professional learning and a
special set of design principles:
multiple voices, self-directed
learning, peer-supported
learning and job-connected
learning.’
Anatoliy Gruzd (@gruzd)
Kellogg, S., & Edelmann, A. (2015). Massively Open Online Course for Educators (MOOC-Ed)
network dataset. British Journal of Educational Technology. http://doi.org/10.1111/bjet.12312
SOCIAL NETWORK ANALYSIS 33

Downloadable from Harvard Dataverse
https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/ZZH3UB

Communication Networks
Online interactions are represented as a graph where
nodes = online participants, and
edges (ties) = communication patterns or other
relation types among participants.

2 Network Files + 2 Node Attribute Files
https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/ZZH3UB
Today’s
focus

Communication Network as a Matrix
USER N19 replied to USER N219

List of Class Participants and Their Attributes

Gephi
Network visualization, data preparation, exploration
Download from
https://gephi.org/users/download/
Requires Java JDK
http://www.oracle.com/technetwork/java/javase/downloads/jdk8-downloads-2133151.html

1. Open File
File > Open
Anatoliy Gruzd (@gruzd)40

2. Run a
layout
algorithm
Try
“Fruchterman
Reingold”
Followed by
“Expansion”

2. Run a
layout
algorithm
Try
“Fruchterman
Reingold”
Followed by
“Expansion”
445 nodes, 1978 edges

3. Identify
Instructors’
network
position
Under “Nodes”,
select the
“Partition” tab and
then “facilitator”
from the drop
down menu, and
click “Apply”

4. Hide
Instructor
nodes
Under “Filters”,
double click
“Attributes” ->
“Equal”;
Drag & drop
“facilitator” to the
Queries section
below and click
“Select”

5. Rerun
“Fruchterman
Reingold”
layout

5. Rerun
“Fruchterman
Reingold”
layout
Anatoliy Gruzd (@gruzd) SOCIAL NETWORK ANALYSIS47
443 nodes, 1450 edges

Explore Research Questions through
Visualizations
What factors influence the formation of communication ties in this network?
Let’s explore the tendency of some nodes to cluster (homophily) and their
network positions (centrality) based on the following attributes:
Connect Whether participants listed
networking/collaboration with others as one of
their course goals on the registration form
Experience2 Number of years teaching
Role Professional role (e.g., teacher, librarian,
administrator)
Grades Works with elementary, middle, and/or high
school students

“Connect”

“Experience”

Gephi Tutorials
https://gephi.org/users/

References
• Steffen-Fluhr, N., Collins, R., Passerini, K., Wu, B., Gruzd, A., Zhu, M., Hiltz,
R. (2012). Leveraging Social Network Data to Support Faculty Mentoring:
Best Practices. Women in Engineering Program Advocates Network
(WEPAN) National Conference, June 25-27, 2012, Columbus, OH., USA.
• Osatuyi, B., Steffen-Fluhr, N., Gruzd, A., and Collins, R. (2010). An
Empirical Investigation of Gender Dynamics and Organizational
Change. The International Journal of Knowledge, Culture and Change
Management 10(3): 23-36. Available
at http://ijm.cgpublisher.com/product/pub.28/prod.1216
• Steffen-Fluhr, N., Gruzd, A., Collins, R. and Osatuyi,B. (2010). N is for
Network: New Tools for Mapping Organizational Change. National
Association of Multicultural Engineering Program Advocates (NAMEPA)/
Women in Engineering Program Advocates Network (WEPAN) 4th Joint
Conference, April 12-14, 2010, Baltimore, Maryland, USA.
Anatoliy Gruzd
(@gruzd)

Introduction to Social Network Analysis

More Related Content

What's hot (20)

Similar to Introduction to Social Network Analysis (20)

More from Toronto Metropolitan University (20)

Recently uploaded (20)

Introduction to Social Network Analysis