Bug? "No Twitter user with that screen name"

Sep 23, 2009 at 4:28 PM


When I use the Twitter search import, on some queries it just stops with the message "No Twitter user with that screen name".  I guess it stops because there's an invalid Twitter user?  Is there any way for the code to skip over these instances and carry on?

Thanks!  Dave.

Sep 23, 2009 at 5:37 PM


One of our team members ran into the same problem, but I was unable to reproduce it.  I should be able to fix it once I can get it to happen myself.

Can you please send me the settings you're using when you run into this?  In other words, which Twitter dialog box does it occur with, and what is entered in the dialog box options?  If you've specified your screen name and password, tell me so and I'll use my own instead.

NodeXL makes several information requests to Twitter.  I suspect what's happening is that some Twitter users have requested privacy, and although Twitter will provide limited information about them, it balks when NodeXL asks for more details.  I can tell NodeXL to skip the details instead of stopping the whole process, but I need to see the problem firsthand before I can make that change.


Sep 24, 2009 at 8:32 AM


I'm using the "Import from Twitter search network" option.

The search is: "true blood" (with quotes)

The "Limit to:" box is NOT checked

The "Include statistics..." box IS checked.

Finally, I do use my Twitter credentials for this job and my username is whitelisted.

Hope that helps!


Sep 24, 2009 at 7:13 PM

Thanks, Dave.  I was able to reproduce and fix the problem.  The fix will be in version

Note that this is one VERY long import.  With the "Limit to" checkbox unchecked and the "Include statistics" checkbox checked, it will get up to 1,500 tweets from Twitter, then request an unlimited list of the people followed by each of those tweeters, then request statistics for each of those tweeters.  Twitter does not have the fastest Web service in the world, so be prepared to wait hours.

-- Tony

Sep 25, 2009 at 3:13 PM

Thanks, Tony - much aprpeciated.

Actually - I have another question.  What search parameters are sent to Twitter apart from the query?  Do you try to retrieve the maximum number of Tweets available through their API for a given query?  Also, in the Twitter search, when you do limit the number of people returned for a query, is this based on the e.g. most recent 500 people (for example) that match that query?  Just trying to gt a handle on what is sent/returned to and from Twitter when the parameters are applied.

Many thanks - Dave.

Sep 25, 2009 at 3:56 PM
Edited Sep 25, 2009 at 4:24 PM

Here is the algorithm used by NodeXL version  The "N" variable is from the "Limit to N people" settings in the dialog box.

1. Ask Twitter for tweets that satisfy the search criteria.  Stop when N unique tweet authors are found.  If N is not set, Twitter stops at approximately 1,500 tweets.

2. For each unique author found, ask Twitter for a list of up to N people whom the author follows.  If the author follows more than N people, Twitter returns the "followed" people in the order in which the author added them as followed.  If N is not set, all followed people are returned.

3. Using the "followed" information, draw edges between the tweet authors.  So if you see authors A and B in the graph and there is an edge between them, it means that both authors tweeted the specified search terms, and A follows B.  (However, if N is set, the lack of an edge between A and B does not necessarily mean that A does not follow B, because the list of A's "followed" may be incomplete.)

4. If the "Include statistics" checkbox is checked, ask Twitter for statistics for each author.

In version, steps 2 and 3 will be optional, and there will be new options for "replies-to" and "mentions" edges.  The new options will use the information obtained in step 1 and will not require additional requests to Twitter.

-- Tony

May 8, 2015 at 11:42 AM
I'm running in the same problem in 2015. Did you find a solution since 2009?
May 8, 2015 at 9:58 PM
Yes, this issue has been resolved. Can you share the configuration that is generating an issue for you?


May 12, 2015 at 3:21 PM
Dear Mark, Ye, sorry for the delay :-( Only now I saw your fast replay. The NodeXL is running in a Desktop Intel Core 2, 64bit, Windows 7, service pack 1, 4GB of RAM, behind the University of Kent firewall. In advance, thank you indeed Daniel deOliveira
Em 2015-05-08 18:58, MarcSmith escreveu:
May 12, 2015 at 3:49 PM
Please contact me directly through CodePlex with an email address or Skype ID. We can schedule a time to review your issue and explore any work arounds.


May 12, 2015 at 4:22 PM
I just noticed the system specifications you posted and I will note that 4GB is a modest amount to work with. NodeXL (and network tools in general) are RAM hungry! We recommend 8GB at minimum for medium sized networks and encourage users to access 16GB of RAM if possible for larger networks.

May 13, 2015 at 1:45 PM
Dear Marc, Once again, thank you for your support. Before your scheduled talk on Skype for this morning, please read the attached document. I wrote a small report of my NodeXL problem, and attached two screenshots. Please see document attached. Thank you
--- Bons codigos / Have a nice code Daniel deOliveira JUG Leader / Founding Java Champion Twitter/Facebook/Skype danieldfjug [email removed] Brasilia Java Users Group www.dfjug.org Brasil Em 2015-05-12 13:22, MarcSmith escreveu: