Marvel Social Networks

Lab Reports, Networks


Like many foreign youths, my initial understanding of American culture was based on Disney animations and Marvel superhero movies. Although I haven’t watched Marvel comics, I have watched all the related movies. Especially the X-Men left a deep impression on me when I was young, and it is also my favorite Marvel series. The X-Men series not only had outstanding production skills of the time, but its success also provided experiences for the subsequent Avengers of Marvel and Justice League of DC superhero movies. So when I saw network data about Marvel characters, I chose this topic for interest. The network visualization presents the relationship between Marvel characters over decades of history and highlights some key characters of the Marvel universe.

Material and Method

The data comes from the Gephi wiki. The Marvel Social Networks of superheroes, constructed by Cesc Rosselló, Ricardo Alberich, and Joe Miro from the University of the Balearic Islands.

The data processing tool used n this project is Gephi, which is open-source software that specializes in network graphs visualization. The readability of data is enhanced by adjusting the layout, filter, and color of the data distribution. “Using Gephi to visualize and understand communities” is a good tutorial about basic functions. Finally, setting some parameters in the preview can further adjust the final appearance of the chart.


Since the number of Marvel characters is pretty large, I first filter nodes with a degree range of 200 or more to remove the very marginal roles, but you can still see that the relationship between characters is very complicated. Adjust the layout mode to ForceAtlas 2, which can gather the more related nodes in the data.

Then I switch on modularity and average degree in the statistic. In the next step I color the nodes according to modularity and the nodes size base on the degree. Using default color selection, you can clearly see several groups, including pink Avengers, green X-Men, orange Fantastic Four, blue Spider-Man, and yellow Asgard.

Next, I adjust the filter to nodes degree range above 900, you can see that the main character in each story are left on the chart. I reduced the thickness of the edges in the preview to make the names of the main characters in each story clearer.

Finally, the minimum degree is adjusted to 1250, and Marvel’s core character is presented. Most of them have their own movies series.

What’s more interesting is that Gephi can also see the relationship between a single nodes and other nodes. For example, if you place the mouse on Iron Man or Wolverine, the characters that have an intersection will be lit.


I think Gephi is a very useful data processing software, and it is also very interesting to process the character relationship diagram of Marvel heros. But the biggest disadvantage of this software is that it cannot undo any operation, which greatly increased the cost of my trials. Because the coloring methods and values of some layouts are not ideal, and the diagram presented in the operation panel will be different from the final state in the preview, it is very inconvenient to re-adjust to the previous state. In addition, in this project, I chose a relatively simple data to process, and it takes more trials and practice to use Gephi to deal more complex network diagrams.