Random sampling networks

Apr 26, 2012 at 3:15 PM

What is the best method to sample a network using NodeXL for Twitter.


I downloaded the complete network (35,000 followers) and compare with a 300 sample I also got from the same network but results does not appear to be equivalent, the 300 sample looks to be biased getting more data of newly joined to twitter users.

Is there a way to do sampling using this plug-in?


Thank you,


Erik Vázquez

Apr 27, 2012 at 5:21 AM


Twitter doesn't say anything* about the order in which it provides information on followers when only some followers are asked for (300 in your case), although I think it's based on when they became followers.  In any case, NodeXL has no control over which followers are provided by Twitter, so the answer to your question is no, you cannot get a random sample of followers.

-- Tony

* https://dev.twitter.com/docs/api/1/get/followers/ids

Apr 27, 2012 at 2:39 PM

This is not the same as asking Twitter to deliver a random sample of followers, but *IF* you collect all followers of a user via the Twitter User Search Importer in NodeXL you could then assign a random value to the Visibility column on the Vertices worksheet.  By setting a threshold for the random value you can control what percentage of the data set will remain visible.