Node Lists and Edge Lists

To build a network diagram, you need a list of nodes (the individual actors) and a list of edges (the relationships between the actors).

Edge Lists

An edge list should have two columns, one labeled Source and the other labeled Target. If your edges are undirected (meaning all of the relationships are reciprocal), then it doesn’t really matter which column is the source and which is the target.

I wanted to build an edge list from the questionnaire you answered. But when I downloaded it from Google Docs, it looked like this:

So how did I get it to look like this?

Easy! I used the Transpose tool on my beloved OpenRefine. Watch:

Node Lists

Really, you don’t even need a node list when you’re using Gephi; it can automatically derive the node list — the list of unique values contained in your edge list. But remember that our node list contains the node-type column, which is what allowed us to divide our nodes into students and preferences. That’s why we uploaded a separate node list.

If you want to upload a node list, all you really need is a list of nodes. But it’s nice to have a column called label, too, so that Gephi knows what label you want to use for each of your nodes.

So how did I get a list of unique values? Easy! I used Excel’s Advanced Filter tool.