IoP Collaboration graph with R igraph

Posted on January 20, 2010 | 6 Comments

My boss asked me to make a nice slide for a presentation showing the collaborations between researchers at our institute. I have a mountain of other work to do, so in the name of procrastination I spent the afternoon playing with the igraph R library:

collaboration
click for big image

Code

This entry was posted in R/Bioconductor. Bookmark the permalink.

6 responses to “IoP Collaboration graph with R igraph”

christian | June 9, 2010 at 1:56 pm | Reply

Hi!

Trying to learn how to use the igraph package in R I came across your beautiful graph. Great, inspiring work!

I tried to do something similar, but sofar I haven’t succeeded. Could you give me hint about the source code of this graph?

All the best,

Christian
cass | July 28, 2010 at 11:56 pm | Reply

Hey Christian – probably a bit late now, but I’ve added a link to the code.
Alex | January 10, 2013 at 3:34 pm | Reply

This is great! Any chance you can let me know the structure of your two data sources. I’m attempting to replicate from a single .csv file. Running into issues when I try to color by data levels.
CassJohnston | January 10, 2013 at 5:05 pm | Reply

Yep. The original .xls file is at http://github.com/cassj/my_bioinfo_scripts/raw/master/collaborations.xls

If you have ruby, rubygems, rake and the Perl module Spreadsheet::ParseExcel installed you can just grab the Rakefile from https://github.com/cassj/my_bioinfo_scripts/blob/master/Rakefile.iop_connections and run ‘rake’. It’ll download all necessary scripts and datafiles, convert xls to csv and run the script to generate the plot for you. I just tried it on RHEL6 and it works for me. I had to make a quick change to the code to get it working with a recent version of igraph but I’ve pushed the change to github.
- Alex | January 10, 2013 at 5:31 pm | Reply
  
  Fantastic. Thanks so much!
Raphaël Pô | September 2, 2020 at 9:55 am | Reply

Hello !
I hope it still alive. I am trying to do the same thing, but to set the edge width proportionally with occurence (meaning in the matrix it can be other figures than 1).
Could it be possible?
Cheers