Using Gephi to Visualise Game of Throne’s

 

There are many visualization tools out there that can enable somebody to see data in a whole new way. The tool I used is called Gephi which allows users to create visualizations using datasets comprised of nodes and edges. I decided to use the tool to visualize the character relationships in a few episodes of the first season of game of thrones to show the characters with the most influence and to visualize how the story is told using a data set I made myself using Microsoft excel. The reason I made a data set in the first place was because Gephi was very specific about the files you imported into the software

Before looking at game of thrones or any topic I inspected different visualization tools. One of the tools I looked at called RAW was capable of taking in different forms data to create visualizations. I did not choose RAW because I felt the visualizations where too simplistic and bland. I choose Gephi because I liked the look of examples I found online and I found the UI offered good opportunities for effective data manipulation. Even though I picked Gephi in the end I had one major issue with using the program and that was actually getting data to use in Gephi. I looked online and found data that was in formats that can be used in Gephi but came with two more problems. The first being that Gephi is specific about what way the data is presented in different formats and the second being that anything I got to work wasn’t really strong data.

I found my best option was to create the data myself using a feature in Gephi to import two spreadsheets in the CSV format comprised of nodes and edges. At this point I decided to visualize the character relationships of the characters in the TV series Game of Thrones. What inspired me to do this was apart from being a fan of the show I thought Gephi best use would be in displaying the relationship of characters in a story. I decided to only visualize the beginning, middle and end episodes of season one as I believe it would give me a brief understanding of the relationships in the world and help me Identify the central character in the beginning middle and end.

In order to make a visualization in Gephi I had to create two spreadsheets one for nodes and one for edges. Nodes are items represented by a number and a label which is a name for the node. An edge is a single connection between a pair of nodes that represents the nodes having a relationship. To make characters interactions into nodes and edges I had to list the characters as nodes and the interactions of each character as edges. To get the information I went to a wiki for game of thrones where you can read the events of every episode. I believed if I gathered the data as I watched the show or read the book it would take too much time. In creating the edges I made them under the following criteria. If character was in a scene with two other characters that character would have an edge made for connecting them with each of the other characters in that scene. The other characters in that scene are connected with each other thus that scene can be represented in Gephi.

The following images are the four visualizations I created in Gephi using a layout known as ‘Fruchterman Reingold’. I kept this layout because I found it to be very presentable with the way the nodes are lay out along with making easy to tell the relationships. To make the information more clear in Gephi I did some actions to make the visualizations easier to understand. The first thing I did was make the node change size based on Average weighted Degree. The degree is a representation of how connected a node is, the greater the degree the node has the bigger it is in my visualization. I also did the same with the color so the nodes will be darker based on degree. I did the same thing with color for the edges and increased the edge thickness to make it easier to interpret.

Game of thrones graphs

 

The visualization labelled “episode one” is the first episode of Game of Thrones where the family known as the Starks are the main focal point of the episode. The Starks take up the most density in the display in terms of node size and color as multiple members of this group are in most of the scenes. The graph also introduces the Lannister’s and the King Robert Baratheon who are central to the episodes plot. The graph also features a group known as the Targaryen’s who are enemies of King Robert in another country which is why they don’t connect.  The three other visualizations are more interesting as multiple groups in all the visualizations are separate from one another.

 

Game of thrones follows different groups who are in different parts of the world during a time where there is political instability. As a result of this it is rare to see central characters interacting with other central Characters; this is shown especially in the bottom two visualizations. The visualization for ‘episode six’ is a good example  of what I am talking about as the four groups shown in the visualization are in for different locations in the world Kings Landing, Winterfell, the Narrow Sea and the Veil. These groups are shown to either to have a lot of interaction among st each other or have little interaction. The visualization does not tell us what the state of the relationships are but do tell us how close the characters are.

The following is combination of all the Nodes and Edges for all four maps using the ‘Fruchterman Reingold’ layout. The second image is the same visualization using the ‘Force Atlas 2’ layout since the visualization in the first image does not give a clear picture of anything other than who is the most dominant character.  The ‘Force atlas 2’ layout sets a distance for nodes based on how closely they are related to each other which is helpful to show my. The visualization shows multiple huge clusters of Nodes separate from one another with only a few nodes connecting these groups. This shows that even though there are many characters huge groups of them don’t communicate with the other groups.

game of thrones full graphwestoros graph

The group in the top right of the visualization are the characters in Winterfell and other places in northern Westeros who are strongly connected. This is due to most of these characters seeing each other as family while in the bottom left there is Kings Landing. In Kings Landing the characters aren’t as strongly connected as those in the north as these characters are both friend and enemy amongst each other. The bottom right represents the veil where a trial is happening and the top right is the edge of the narrow sea where an army is being built to take Westeros.

These visualization help solve one question that people have about game of thrones which is who is the central character? Based on my visualizations the most influential character in season one is Eddard Stark who is the hand of the king. Eddard Stark as hand of the king does most of the king’s work which includes making decisions and holding meetings while being responsible for advising the king. In reality it is a more powerful role than the king, in having such a powerful role Eddard is made the central character in the plot which is why in my visualization he has the strongest node. In my visualization for episode ten Eddard Stark is not present yet is still dominant when I put all the visualizations together.

My Datasets:

Nodes – GOT all nodes

Edges – GOT all edges

References:

Gephi – https://gephi.org/

RAW – http://raw.densitydesign.org/

Game Of Thrones wikia   – http://gameofthrones.wikia.com/wiki/Game_of_Thrones_Wiki

 

 

 

 

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s