“Gephi is an open-source software for visualizing and analysing large networks graphs. Gephi uses a 3D render engine to display graphs in real-time and speed up the exploration. You can use it to explore, analyse, spatialise, filter, cluterize, manipulate and export all types of graphs.”
Gephi is a program downloaded from the internet, and can be found here.
First your group should decide what instances/variable/interactions you will be collecting information about. This could be who talks to who in the novel, who directly interacts with each other, or who as a character is given direct speech rather than internal dialogue. Whether or not these interactions are directed or undirected becomes important when you create the edges within the Data Lab, and it will be easier to keep the two edges separate so you know who directly interacts with each other.
For example, Character A initiates a conversation with Character B, making this “edge” a directed one. You’ll be able to make your own distinctions about what network you are plotting, but it helps to have an concrete idea so that you don’t lose track of what your network is illustrating.
Once you have your data collected you can put it in an Excel Spreadsheet so that you can keep the data incase you ever need to use it, or alter the data that is being placed into the Gephi Data Lab.
Open Microsoft Excel:
You will need two separate spreadsheets: one for Nodes and one for Edges.
In order to get spreadsheets into Gephi they must be saved as a .csv file.
Nodes columns: Id, Label, and Modularity Class
Id: The number assigned to your character.
Label: The character you are collecting data about.
Modularity Class: The number of appearances within a text/chapter/given length of time–the number you are keeping track of.
Edges columns: Source, Target, Type, Id, Label, Weight
Source: The ID from the nodes list (the character who initiated the interaction)
Target: The ID from the nodes list (the character who is participating in the interaction, but not the origin of the interaction.)
Example: If Character A talks to Character B, Character A is the source node, and Character B is the target node
Type: Was the interaction directed/undirected. Directed could mean that there was an active action, and is determined by what you are recording: Character A talks to Character B; if we were keeping track of how often characters talked to each other–then this would be a directed interaction. Undirected creates a solid line between two nodes, while directed creates an arrow coming from the source node to the target node.
Id: The number assigned to your edge (begin at zero.)
Label: If you’d like to assign text to your edge–you’d put that here–though it must be unique.
Weight: The number of times a character interacted with another. If Character A talks to Character B forty times within a given timeframe, you would use this number. However, if you are keeping track of who initiates conversation (directed relationship) and Character A initiates conversation with Character B ten of those times, then you would have to create another row where Character B is the source, Character A is the target, the Type would be directed, and the weight would be thirty, for the other times that are left unaccounted for. The width of the line is determined by this number.
Save both spreadsheets!
Open Gephi:
Once all of the data is in your spreadsheet open a new Gephi project. Click on the Tab called “Data Laboratory;” then click the tab that says “Import Spreadsheet;” a menu will pop up that has a search bar for you to browse your computer to locate the .csv file with your data. Within the browser window make sure that the File Format is set to “CSV/ *csv.”
**Overview Panel is where you can manipulate how the Preview Panel will look. It would probably be most helpful to use this other Gephi Tutorial when first starting out.
We now need to import the Nodes List:
Make sure your settings look like this:
These settings are default, but just to make sure the drop down menus should read as: Separator: “Comma;” As table: “Nodes table;” Charset: “UTF-8.” In the Preview window, your data should look like what you’ve saved in the spreadsheet. If all of this is correct, click “Next.” Import settings: Imported columns (the Id box should be checked, and set to string; the Label box should be checked, and set to string, and “Force nodes to be created as new ones” should also be checked. “Finish” You will now see your data in the Gephi Data Laboratory. (Below)
Now, we’ll add the Edges List:
Make sure your settings look like this:
Okay, now that you have your data in this panel, it will appear on the Overview Panel:
A quick overview on the Overview:
To make your character network easier to read you’ll want to work first in the Overview tab, There will be multiple sectioned screens like “Partition,” “Rank,” “Layout,” “Statistics,” and “Filters.” To highlight the number of interactions between characters you can change the “weight” of the nodes and edges within the “Rank” tab. Again, this Gephi Tutorial will be very helpful!
Adjust the nodes on the Overview screen so they are located where you want them, now you can start adjusting the colors of edges and nodes within the Preview Tab.
Preview
Check the Preview window to view your progress, and remember to click “refresh.”
To show Node Labels check the box to the left to make them appear. Then press “Refresh.”
There will be multiple compressed and expanded menus reading “Nodes,” “Node Labels,” “Edges,” “Edge Labels,” and “Edge Arrows.” adjust the settings to your liking, and remember to hit the “Refresh” button at the bottom of the screen. Remember to save the settings when you’re all finished.