2. Data Source:
• Social collaboration network of characters in
the Marvel Universe
http://syntagmatic.github.io/exposedata/mar
vel/data/hero-network.csv
• Nodes: Marvel Characters (10469)
• Edges: when characters are in the same
comics (178115)
• Undirected network
• Not weighted
Dataset & Visualization Characteristics
Tool used: Gephi
• Provides rich UI for graph
visualization
• Adjusted node size and colour to
show degree centrality
• Applied ForceAtlas2 layout function
• Modularity to show communities
• Additional metrics for other measures
Objective: Use Graph / Network Mining technique to do the below ones
• To determine the main communities and the key player in Marvel Universe
• To determine the main key player whose removal would destroy Marvel Universe
• To predict main key player that would be added to upcoming Avengers movie
3. Origin and History
• Born in October 1939 & by 1950’s, had generally
known as Atlas Comics
• The Marvel branding began 1961, the year that the
company launched The Fantastic Four and other
superhero titles created by Stan Lee, Jack Kirby, Steve
Ditko, and many others
• Most of Marvel's fictional characters operate in a
single reality known as the Marvel Universe
• Marvel Ecosystem – 30000 comics, 5000 writers and
8000 characters, 46 movies and TV shows, 100+ video
games
• 40 years of animation, toys and consumer
products
4. Community Detection & Key players
• The Marvel dataset is composed of a list of co-occurrences of super
heroes / characters in comics /movies
For example, every time Spider Man appears in a comic book
with Captain America, we will have a line with both their names
• The nodes with bigger size corresponds to the key player
• Node colour detects the various communities
• The heroes with higher degrees are Iron Man, Captain America &
Spider Man
• 25 Communities using Modularity metric (seen in next slide)
• Modularity - 0.49
6. Community Detection
• After increasing
the degree 600
• 6 prominent
Communities
captured
Key players:
• Nodes with
greater size
7. Closeness Centrality
• Closeness Centrality and Yifan Hu algorithm is applied to
this network. It distances the communities that are
independent to the network.
• In the adjacent figure, communities
which are found outside the highlighted
dotted lines are independent where the
actors are not related anyone in the
bigger network.
• Heros Reborn remnants which is an
one off comic book released by
MARVEL, and the characters are not
been re-appeared in any of the other
comics/movies.
• There are many such nodes found in
the network inferring that those
characters were not become famous, so
not introduced in any of the comics.
CLOSENESS CENTRALITY
8. • Shows which nodes are connected to
other nodes that have high degree. It’s a
measure of influence of a node in a
network.
• Highest influencer in the network is
Captain America followed by Spider-man.
• Captain America acts as the authority
Eigen Vector Centrality
9. highlow
Nodes with high betweenness
may have a considerable
influence within a network
by virtue of their control over
the information passed
between the other nodes.
Betweenness node removal
causes a disruption in the
network
Spider-Man & Captain-
America have the highest
Betweenness Centrality
among all other nodes
Betweenness Centrality
10. • 4-clique. 4-Clique implies that sets of
characters appeared together together in
comics. Also, the fact Mr. Fantastic lead as
a central role for other three actors.
• Richards Reed (Mr. Fantastic) plays a
central role in the lives of his wife
(Invisible Women), hot-headed younger
brother Johnny Storm, and Ben Grimm,
her close friend.
• This clearly shows a strong bonding
relationship formed a clique.
• Sue Storm became a more powerful
member of the Fantastic Four, and the
team's second-in-command with a
growing assertive confidence. While Sue
operated somewhat in the shadow of her
brother and her husband in the early
years, she is now the soul of the Fantastic
Four and one of the primary heroes in
the Marvel Universe.
4- CLIQUE - FANTASTIC FOUR
11. EVOLUTION OF MCU ( Marvel Cinematic Universe)
Marvel made relatively little profit from its licensing deals with
other studios and wanted to get more money out of its films.
The below characters has been licensed to other production house, ddd
• Spider Man – Sony Pictures
• X Men series – Fox Studios
The idea of shared universe – Comic popularity
Launch pad – Introduction of Individual nodes ( Individual
Character film) and forming a betweenness Centrality of all nodes
( Film comprising all characters ).
How they used the ecosystem efficiently and made a profit out of
that?
INITIALIZING AVENGERS
12. Insights – Avengers
Betweenness
Centrality*
• Main key players are already in Avengers team
• Wolverine being key player is no more, so may not make
any appearance in Avengers
• Spider Man is expected (predicted through Graph mining metric) to
join in upcoming Avengers Movie and he is one who can link avengers
team due to high betweeness centrality.
Eigen Vector
Centrality