Twitter as a data shed

Please use the two following two hashtags when commenting on this post on Twitter: #blc11 #datashed

See update: Twitter as label-maker

Brian Mull at November Learning and I had a brief email exchange yesterday on how to develop online conversations around data points, in this case, videos. I don’t like the sort of comment threads that run down in a long column at the end of a blog post. They’re difficult to scan, harder to search and quickly become a tangle of discussion threads.

That got me to thinking about another conversation I had way back in 2008 with forward-thinking friends, Peter Rawsthorne and John Dumbrille. We were exploring the idea of using Twitter as a data storage platform.

     As I mentioned in my presentation on the Digital Learning Farm at BLC 11, my students and I at Think Global School had a terrific tool for storing and working with large amounts of unstructured data. Spot, as it’s called, lets users upload and tag just about any sort of data: text, blogs, images, videos, bookmarks, discussion threads. The data can then be sorted on the fly in any number of possible mashups made by combining two or more tags in a search.

     There are a number of advantages to storing and using data this way. First, it encourages students to think conceptually about discrete pieces of data in order to tag it. Secondly, it encourages students to build connections between seemingly disparate sets of data and similarly encourages interdisciplinary work. For example, I asked my math teacher to have her students tag their scale models of the Great Wall of China with the same tag my literature students were using in their study of 8th century Chinese poetry featuring the Wall. Most importantly, it encourages creative thinking with data; the user can mash things up in endless ways–like doodling with data.

     But, I now face two problems. I’ve left TGS to become the Director of Educational Technologies at Mulgrave School in Vancouver, so I don’t have access to Spot anymore. (TGS plans to release Spot as an open source platform sometime in the future.) But even if I did, it’s something of a walled garden right now. Though TGS is working on a way to allow broader collaboration, you currently need a TGS email to use it and that restricts the user base to an in-house crowd.

     I’m looking for a tool or set of tools open to everyone that will duplicate at least some of the functions of Spot. Right now.

     Here’s an alpha workaround. It uses Twitter as a sort of label maker and data storage shed. I say shed because it’s the kind of place where you open the door and check things in. Here’s how I think I can recreate the data mashups I made in Spot:

  1. Post a discussion topic, video, assignment on a blog or some other public platform.
  2. Clearly indicated in the post which two (or more) hashtags users must add to a tweet in order to participate in the discussion
  3. Set up a saved search for the hashtag combination in your Twitter client so you get a more or less real time feed for the data
  4. Tweet away
  5. When 140 characters won’t do, post to a blog, YouTube etc. and link to it in a tweet with the hashtag combination.

I’m fascinated with the way this system stores data in the ether, which seems to me to be a little beyond the cloud even. Sure it sits on Twitter’s servers, but it doesn’t feel like it’s in any particular location like a Dropbox folder or in a Google Apps collection. Indeed, the data is, as I said, unstructured until the user gives it form with a search order.

In my conversation with @prawsthorne back in ’08, I said

Experiment with tagging web content. The work that Peter Rawsthorne and John Dumbrille are doing on tagging web content has tweaked my interest as a teacher. I really like the way this is not platform dependent and would be easily transferred to work on a cell phone–which is where I think things are going to go in the classroom. This may transcend things such as Delicious and Diigo

 And I still like this idea for the same reasons. Twitter is simple. It’s open to everyone so there’s no need for managing membership like there is with Spot or delicious or Diigo. It’s and platform agnostic and, as I said, easily accessible from a smart phone. It has a huge user base which encourages breadth of data gathering and discussion. And one feature I especially like is that users pull the data they need instead of having it pushed out to them. You can even work round the 140-character limit and take the conversation deep or add video or image in a blog or any other public platform and tweet a link to it, making sure you add the agreed upon hashtags. 

The downside? Plenty: 140 characters is limiting. Hashtags and twitter handles will reduce the character count further so this will restrict how many tags can be added to a tweet. As it is, I think two hashtags are a necessary minimum in order to focus the data stream and filter out most extraneous data. (But misspell a hashtag and the data is lost.) This is wide open to spamming, too. Finally, in the case of linked blogs, images or video, the hashtag often isn’t on the object itself. I don’t know if this is a real issue, or just my anxiety over really busting out of the old filing cabinet way of thinking. But if I think of Twitter as a label-maker and sticking tweets onto objects, this isn’t so unfamiliar.

At any rate, here’s my first shot at this:

Screen_shot_2011-08-04_at_8

You can see the same by saving a search for #blc11 #datashed in your Twitter client. It’d be great if you add comments on twitter using the hashtags just so we can see if this works, at least in principle. I’ve found a another problem that needs solving: getting past the 140-character limit so we can add more tags. More tags means more ways to structure the data and better ways to organize conversation threads. Google+ might solve this–here are the thoughts of Chris Messina, the fellow who brought tagging to Twitter.

Thought, comments are welcome, especially if they’re on twitter.

Leave a Reply