Creating a network from a table of entities and their attributes


last modified: 2017-02-23

== !

gephi logo 2010 transparent

Presentation of the plugin

This plugin is created by Clement Levallois.

It converts a spreadsheet or a csv file into a network.

This plugin enables you to:

  • Start from a data table in Excel or csv format

  • In the data table, nodes are the entities listed in column A

  • Nodes' attributes must be listed in columns B, C, D, etc.

  • Connections will be created between nodes, when they have identical attributes.

  • Attributes can have values, stored in columns right next to the attribute.

1. The input

An Excel file
Figure 1. An Excel file

2. The output

Figure 2. Resulting network

Installing the plugin

Choose the menu Tools then Plugins
Figure 3. Choose the menu Tools then Plugins
Click on the tab Available Plugins
Figure 4. Click on the tab Available Plugins
Install the plugin then restart Gephi
Figure 5. Install the plugin then restart Gephi

Opening the plugin

Open the plugin via the menu File   Import
Figure 6. Open the plugin via the menu File - Import

Using the plugin

First panel

Select a file
Figure 7. Select a file
file without header en
Figure 8. A file without headers
file with header en
Figure 9. A file with headers

Second panel

plugin 4 en
Figure 10. Parameter for weight

Third panel

plugin 5 en
Figure 11. Confirmation panel

How is the similarity computed, exactly?

We use the cosine similarity. Sounds complicated, but it is not. Check here.

The source code for the cosine calculation is in this file, at this place.

FAQ / special notes on the plugin

1. Excel files should be .xlsx, not .xls

Because they represent two slightly different files formats, and the plugin supports only .xlsx

2. csv files are ok.

If you select a csv file, you will be asked to indicate the field delimiter and optionally the text delimiter.

plugin 6 en
Figure 12. When a csv file is selected

3. You can’t use numerical values in the attributes

numerical attributes en
Figure 13. Age is a numerical attribute

This is too bad. If there is enough demand for it I’ll add this feature, which is not trivial.

4. Each entity should appear only on one line

plugin 7 en
Figure 14. An entity appearing twice

David appears on lines 2 and 5 (because he made two purchases). Only the latest line where David appears (line 5) will be taken into account.

The end

Visit the Gephi group on Facebook to get help,

or visit the website for more tutorials