DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING
COLLEGE OF ENGINEERING, GUINDY,
ANNA UNIVERSITY, CHENNAI – 600 025.
CS6029 – SOCIAL NETWORK ANALYSIS
CREDIBILITY ASSESSMENT OF USERS ON
QUESTION AND ANSWER PLATFORMS
Team 5:
Shreyas Ramanathan A – 2022103527
Jaasim Hameed S – 2022103532
Sri Rama Pandian H U - 2022103537
1. Tool Brief
1. Graphviz
Main Features:
● Open-source graph visualization software.
● Uses the DOT language to describe nodes and edges.
● Offers multiple automatic layout engines (hierarchical, radial, force-
directed).
● Highly customizable: node/edge colors, shapes, labels.
● Can be integrated into programming workflows via Python (PyGraphviz)
and R (Rgraphviz).
Benefits for Social Network Visualization:
● Great for structural clarity in networks (especially organizational
charts, information flows).
● Handles small to medium-sized networks well.
● Useful for academic research and conceptualizing complex
relationships clearly.
2. Sprout Social
Main Features:
● Primarily a social media management platform (not pure SNA
software).
● Provides visual analytics for interactions, audience engagement, and
messaging.
● Offers network graphs of how users engage with your brand across
platforms like Twitter, Facebook, and Instagram.
● Strong reporting and dashboard features.
Benefits for Social Network Visualization:
● Helps brands and social media managers visualize audience behavior
and interaction patterns.
● Useful for tracking influencer networks, customer engagement hubs,
and message spread.
● Good for real-time insights into how information flows through digital
communities.
3. Gephi
Main Features:
Open-source software for network analysis and visualization.
Interactive and dynamic visualization with a focus on large-scale
networks.
Supports graph manipulation via the Gephi Toolkit for programmatic
control (Java).
Extensive layout algorithms, including ForceAtlas2, Fruchterman-
Reingold, and Yifan Hu for layout optimization.
Real-time graph manipulation with drag-and-drop functionality.
Allows dynamic graph creation and manipulation with temporal data.
Network statistics and metrics: centrality, modularity, clustering
coefficients, etc.
Supports filtering based on graph properties (e.g., node degree,
edge weight).
Benefits for Social Network Visualization:
Great for large-scale networks: Gephi excels at handling large
graphs (thousands of nodes and edges) efficiently.
Provides a user-friendly GUI for visualizing and analyzing networks
without needing to write code.
Real-time manipulation: You can interactively adjust the graph
layout, zoom in/out, and highlight specific parts of the network.
Visualization flexibility: It offers highly customizable visualizations,
allowing you to change node/edge color, size, and shape based on
metrics like centrality, degree, etc.
Analysis tools: Gephi allows in-depth analysis of network structure,
including finding communities, measuring centrality, and
understanding flow dynamics.
Suitable for various types of social networks: It’s great for social
media analysis, organizational structures, and collaboration
networks.
Wide range of exports: Can export to formats like PDF, PNG, SVG, and
also network analysis data formats like GEXF, CSV, and GraphML.
2. Matrix representation
The network created for this project is a directed graph where each node
represents a user. Edge from A → B means "A answered a question asked by B.".
Weight on the edge represents number of times A answered questions of B.
Type of Relationship Meaning
Edge from A → B A answered a question that B asked
Edge weight How many times A answered B's questions
Type of Relationship Meaning
Out-degree of A How active A is in helping others
In-degree of B How much help B receives from others
Due the large number of users in the dataset, a sparse matrix was created using
Python modules.
Topic:4
The first 10 rows and columns are printed for representation purposes.
The matrix is visualized using a sparse plot. The X and Y axes represent indices of
the matrix. Points represent entries of the matrix.
3. Node-link diagram
The node-link diagrams were created using the top 100 nodes with highest
degrees and edges between them.
The above network was visualized using Graphviz. Graphviz is a command line
application for network visualization. It involves creating a .dot file with the edges
and running the command to generate the node link-diagram.
The above network was visualized using Gephi. Gephi allows for more complex
visualization than tools like Graphviz. In the above graph, nodes with higher
degrees are represented using darker shades of green. Similarly, edges with higher
weights are given darker shades.
4. Applications Discussion
1. Cover Networks
Definition:
Cover networks map relationships based on who covers or reports on a particular
event, person, or topic — often seen in journalism, media studies, or online content
tracking.
Applications:
● Media Analysis: Tracking how news stories spread across different
outlets and identifying key news hubs.
● Misinformation Studies: Understanding how false information proliferates
across news websites and social media.
● Brand Monitoring: Companies analyze cover networks to see
which influencers or media outlets are amplifying their brand.
● Crisis Communication: Governments and organizations map coverage
during crises (e.g., natural disasters) to optimize public information
flow.
Significance:
They reveal information diffusion patterns, influence hierarchies, and agenda-
setting mechanisms in media ecosystems.
2. Collaboration Networks
Definition:
Collaboration networks represent the relationships between individuals or
organizations that work together on common projects (e.g., researchers co-
authoring papers, developers building software, companies forming partnerships).
Applications:
● Scientific Research: Mapping co-authorships to study knowledge
production, innovation hotspots, and influential researchers.
● Business Innovation: Understanding how firms collaborate for R&D
and technology development (e.g., biotech partnerships).
● Creative Industries: Visualizing collaborations between artists,
musicians, filmmakers to explore creative clusters.
● Open-Source Software: Analyzing how developers contribute to shared
projects on platforms like GitHub.
Significance:
They highlight knowledge sharing, resource pooling, and innovation dynamics
across disciplines and industries.
3. Co-Citation Networks
Definition:
Co-citation networks connect documents (or authors) that are cited together in
later works. If two papers are often cited in the same context, they are linked.
Applications:
● Literature Mapping: Identifying major schools of thought and emerging
research areas in academic fields.
● Bibliometric Studies: Understanding the intellectual structure of
disciplines by analyzing how ideas are connected.
● Recommendation Systems: Improving scholarly search engines and
recommendation algorithms (e.g., suggesting related articles).
● Patent Analysis: Mapping technological fields based on patents
cited together.
Significance:
They reveal thematic clusters, historical development of knowledge, and
influential ideas shaping a field over time.
5. Co-Authorship Network Visualization
Dataset source : http://snap.stanford.edu/data/com-DBLP.html
The bibliography network for computer science was obtained from the
above source.
Nodes represent authors and edges between authors implies they have authored
atleast one paper together. The dataset contained the edges of the network in
COM-DBLP files. For better visualization, a subset of 500 rows was considered
from the extremely large dataset. They were converted to CSV files using
Visual
Studio Code editor and used for visualization. The visualization tools are Graphviz and
Gephi.
Gephi visualization
Graphviz visualization
