Number of Tweets from Twitter Search Network and Turkish Characters

Dec 23, 2013 at 8:23 PM
Hi all,

I've been trying to find all the twitter data around a hashtag. I use import/from twitter search network and enter the hashtag but I only get limited number of data such as 400 (eventough I limit to 10.000) and only from last 6-7 minutes. For example data relationshipdate It starts with GMT time 20:00 and ends in 20:07. Is there any way to exceed this data limit and reach data from yesterday?

Another issue is I can't search for words including Turkish Characters such as ş, ı, ç etc. Is there any way to hande this?

Thanks
Coordinator
Dec 23, 2013 at 8:38 PM
A workaround for non-English characters is described here: http://nodexl.codeplex.com/discussions/451206

-- Marc
Coordinator
Dec 23, 2013 at 8:50 PM
Edited Dec 23, 2013 at 8:51 PM
Twitter delivers data to the public API in unknown and possibly unknowable ways. Often the greater the volume of tweets on a topic, the fewer the tweets it will deliver. For slower moving topics Twitter will often deliver more data than fast moving data. Which tweets are dropped? Good question! The answer might change by the minute as Twitter tunes its various resources. Of course, there is the added issue that Twitter will not deliver data aged more than about a week to (occasionally) 10 days. Access to archival Twitter and other social media data is now a business. See folks like Topsy (now a part of Apple), DataSift, PeopleBrowsr, GNIP, and others like Radian6 (Salesforce), Crimson Hexagon, and others who resell the Tweet stream (among other social media data feeds). Sadly, these are not inexpensive products. NodeXL attempts to perform as best as possible for a client operating within the public API. The public API does not deliver the volume of data that the paid services do (and even those often do not guarantee an exhaustive collection).

It is possible to automate the operation of the NodeXL data collector and to repeat collection every hour. The result may be many fragmentary slices of the data, but it is likely to generate more data than just one or a few manually performed collections.

See: How to schedule the creation of a network with NodeXL and Windows Task Scheduler
http://www.connectedaction.net/2010/08/23/how-to-schedule-the-creation-of-a-network-with-nodexl-and-windows-task-scheduler/

-- Marc