Gephi Workshop 1

Copyright 2011 Digital Enterprise Research Institute. All rights reserved.

Digital Enterprise Research Institute www.deri.ie

Enabling Networked Knowledge

Gephi Workshop 1

David CrowleyMaciej Dabrowski



Open Gephi On the lab machines it might have to “install” or

“download” it but should only take 2-3 mins

We are going to play with the “Les-Miserables.gexf” sample file to start

Save a copy so you can always start again with the original file



Change to Directed





Les Miserables Network

Co-appearance weighted network of characters in the novel “Les Miserables” from Victor Hugo.

Nodes are placed randomly so it may look different for you



Controls



Layouts

Layout algorithms sets the graph shape, it is the most “essential action” (for visualising)

Graphs are usually layouted with “Force-based” algorithms. Their principle is easy, linked nodes attract each other and non-linked nodes are pushed apart

Layout Panel Shows layout choice And Layout Properties



Force Atlas 2

Run the layout for a few seconds and press stop

You should end up with a graph like this (not exactly the same or it could even be quite different)



Layout Properties

The purpose of Layout Properties is to let you control the algorithm in order to make a aesthetically pleasing representation

Change Scaling (repulsion) to a high number and see what happens and then make it small

Each time we need to run and stop the layout before any visible change happens.



Layout with Scaling @ 50



Ranking

Choose Ranking on the Right Hand Side panel above the layout panel

Nodes should be selected

In the drop down menu you should have three options Degree Indegree Outdegree



Ranking

Choose Degree Click Apply

Now your graph should be coloured in

You can change the colour settings by clicking on the colour bar



Result List

Click on Result List as highlighted by the red arrow

Click Apply

Now you should see the nodes listed by Rank



Metrics

We will calculate the average path length for the network.

It computes the path length for all possibles pairs of nodes and give information about how nodes are close from each other



Choose Directed – Click OK



Results



Ranking

In the Ranking Panel now in Nodes, we have more metrics

Every (or almost every time) you run a metric you add a new way to rank the graph



Ranking (size) with overlap turned off



Community Detection (Modularity)

Community Detection – used a lot!

Can we predict collapse of online communities?

Concept of hubs – key points in the network

Crime networks – old studies on white collar crime



Modularity

Run modularity

Now you can rank nodes by modularity

This should identify “communities”



Partition

If you click on Partition (next to Ranking)

Press Refresh

You can partition by Modularity Class



Filters

You can run Filters on your graph

Useful for large graphs

Finds components or hides components

Gephi Workshop 1

Documents

layout panel nodes

nonlinked nodes

possibles pairs of nodes

graph shape

visible change

average path length

colour settings

forcebased algorithms