Tuesday, March 10, 2009

Mining the tweetstream

Last night's inaugural educhat conversation on twitter prompted me to again think about how we might mine the tweetstream for information on teaching and learning. I collected all tweets tagged #educhat (thanks to the help of others, such as @aforgrave), and also did this recently for the tweets tagged #colearning from Colorado Learning 2.0. This morning while driving from place to place, I recorded some of my thoughts (via gcast) on mining and coding this data. I'd be very interested on your input and ideas- please listen and comment, email me, or contact me on twitter (@Bud_T). 

Subscribe Free
Add to my Page


  1. I find twitter difficult to use with tagging. I don't think my mind works well that way. I think that tweetdeck makes it easy to do that, it just doesn't seem to work for me. It could be that I don't look at Twitter as filling that particular need for me.

  2. I'm guessing a simple exploratory study would be a good first step. Just looking to identify a subset of Tweets & categorize by topic discussed / posted. Also, looking at dialogue vs. posting -- how many tweets are posting resources vs. actual conversation?

    The hardest part is getting a meaningful sample. Let me know if you want to pursue this. I'm interested.

  3. I'll probably pursue this sometime after April(conferences call, and papers need to be written). I agree that the sample is the issue. Perhaps after collecting a few educhat's worth of tweets, a decent sampling scheme can be thought out.