The Marvel Social Network Analysis



Since 1939 Martin Goodman created Timely Marvel Comics, the Marvel brand used over the years. Marvel creates numerous characters, such as well-known superheroes: Spider-Man, Iron Man, Captain America, etc. Also, Superhero teams exist like the Avengers, the X-Men, the Fantastic Four, and the Guardians of the Galaxy. All ther fictional characters in Marvel world operate in a single reality known as the Marvel Universe, with most locations mirroring real-life places. The majority of characters are based in New York City. As the movie industry revolution through 3D effect technology, commercial films inspired by the Marvel story and produces serious superhero movies. In this generation, many audiences go to movie theaters to watching hero movies to support the dream comic books from their childhood. They also enjoy the sound and light effects from the big screen and follow the novel trend in the animation industry.

Material and Method

In this project, I used the data visualization tool, Gephi, which is open-source software that specializes in network graphs visualization. To learn how to use Gephi, I followed the online tutorial “GEPHI – Network visualization tutorial [HD]”, which created by a data scientist, Martin Grandjean. In this video, Martin presented how to enhance data and graphs by adjusting the layout, filter, and color of the data distribution.

Also, I explored the dataset from the Gephi wiki on Github, which is an open-source contributed by multiple cooperation partners. On this page, I downloaded the Gephi file regarding my topic, The Marvel Social Network.  This dataset of superheroes network is constructed by Cesc Rossell√≥, Ricardo Alberich, and Joe Miro from the University of the Balearic Islands. Collected by Infochimps and transformed & enhanced by Kai Chang.

Analysis Process

In ther first step, I imported the dataset into Gephi and found out the number of Marvel characters is enormous and complicated. In this graphic (Figure 1), the social network of superheroes is hard to read, and the scope of information is widely but not so useful. To see the name of the character clearly, I tried to zoom in on the picture (Figure 2) and figured out the details of which part of this dataset is my target material that related to my topic.

Figure 1: Screenshot from importing dataset in Gephi
Figure 2: Screenshot from Gephi

First, I try to change the color since the back-and-white is not quite readable (Figure 3). I picked up the three colors (Blue, yellow, and red) to show the graphic. The main character, such as Captain America and Invisible Woman, would present in the primary orange color. And the characters who are not so important would be shown in the blue color.

Figure 3: Screenshot from Gephi (Degree Range of 200)

Then, I decided to filter the initial filter nodes with a degree range of 200 (Figure 4). Also, I intend to remove the roles that are not so important in the Marvel story. And I keep trying to narrow down the scope of the dataset with a degree range of 100, 50. (Figure 4, 5)

Figure 4: Screenshot from Gephi (Degree Range of 100)
Figure 4: Screenshot from Gephi (Degree Range of 50)

However, the network graphic is still not clear enough to read the connection between each character. I decided to focus on the Giant Component and adjust the filter with the degree range 20 and get the result (Figure 5). In this picture, I can see the name of the main characters and their network clearly.

Figure 4: Screenshot from Gephi (Giant Component with Degree Range of 20)


Finally, I got the conclusion of this picture. I adjust the gradient color with the spline curve to enhance the three colors. To make the graphic more readable, I also replace the default black color text with their parent color. And I colored the nodes according to their degree level and resized the node border range. With the different color selections, the picture clearly shows each category and group connect with the main character, Professor X.


The Gephi is a very useful data processing software, but the loading process is quite laggy. I had a hard time rendering the picture and zoom in and out since the system running slowly. However, it’s very interesting to see the network between each character and learn about their relationship diagram in Marvel story. Also, I found out the software cannot save the working history and redo the previous step. For example, I cannot compare the 200-degree range filter and the 50-degree range filter when I adjust the Giant Component. And every time I rendered the picture again the software seems stock, running very slowly, and even shut down my computer without saving the result.


The article, Marvel Comics from Wikipedia.

The tutorial, “GEPHI – Network visualization tutorial [HD],” created by Martin Grandjean.

The dataset, The Marvel Social Network, contributed on Github.