I see Tweets as important texts for analyzing the dynamics of fandom.
But getting Twitter content into analyzable form is a major obstacle in my research. I’ve discovered a few tools for converting Tweets into plain text:
http://listoftweets.com/ : This site allows you to build a text out of multiple twitter searches, but it limits each search to the 100 most recent tweets with that hashtag.
http://tweetkeeperapp.com/ : This app looks promising for creating text archives of particular twitter users, but it doesn’t collect tweets based on hashtags. It allows you to export your archives in plain text.
http://mashe.hawksey.info/2012/01/twitter-archive-tagsv3/ : This tool allows you to download a google spreadsheet and set your own parameters for collecting tweets. I used this tool in the Summer to archive the twitter backchannel of a conference that I attended. I can’t find a direct way to turn the archive into mineable text, but I could theoretically copy and past all of the tweets into a doc and then delete the image objects.
(for some reason, I’m not able to change the display text on the above links–sorry about that!)
Okay, so assuming that I use some combination of the above tools (and/or others that I have yet to discover), the text that I would like to analyze would be a long string of tweets concerning a particular television show.
I obviously need BYOT kinds of tools for this operation, but without a really solid text to test out, my search was a bit frustrating.
Some of the tools that I encountered were pretty confusing (Gephi, Mallet). MONK only analyzes what they have in their library–I think. Several of the tools require you to download software (Gephi, R, Mathematica).
Voyant seems the most user friendly easiest to use, since you can just copy and paste your text. I tested out a document I created using “List of Tweets,” but I don’t think it was long enough to generate anything interesting. I’m going to have to do more work to establish my TwitterText, then give these tools another go.
Looking forward to talking about everyone’s findings tomorrow!
Cheers,
Casey
My guess is that you’ll need to think about building some custom tools: existing ones for Twitter tend to be limited to what the Twitter API allows for searching, and generic tools aren’t designed for many small documents with particular characteristics (@ and # symbols, retweets, etc.). Are you up to learn some coding?