Terrell Russell: This Old Network » cloudalicious

Reflections on iConference08 at UCLA

Terrell Russell — Mon, 14 Apr 2008 04:06:33 +0000

Over a month ago (how did that happen already?), I attended the iConference at UCLA and had a great time. I spent three nights and four days in a blur of activity and ideas. Over my time in Los Angeles, I visited Kinko’s twice, generated three presentations, gave three presentations, and had fancy finger-foods at the Getty Center (and Research Institute and Museum) (which has one of the most stunning architectural layouts I’ve ever seen).

Doctoral Colloquium
I spoke about my ongoing progress towards my dissertation topic of Contextual Authority Tagging and laid out my most recent plan of attack (pdf 736kB). I can see how the pieces are fitting together now and did my best to convey a few years of work into 10 minutes. It’s not an easy thing to do, but I’m getting better at it. Practice definitely helps.

Poster Session
The next day, I shared my newest results concerning Cloudalicio.us and its use for seeing the tags used on a group of items change over time. This was the first time in public for this data and these views, and I received some wonderful feedback regarding periodicity (and the potential predictive power of these graphs) as well as generalizability.

Watching Organizational Opinion via Social Tagging (pdf 424kB)

It turns out, if I can describe the type of data I’m graphing better – others may be able to push their data into Cloudalicio.us and see how their own data is changing over time. This is very exciting as it opens up many potential collaborations – with people and datasets I’d otherwise not have an opportunity to see.

Panel Session
The third day of my whirlwind week was an early-morning presentation of the contribution that may have the most impact on others doing research. Tag Decay (pdf 176kB). I posit that by adding time to the tagging ‘triumvirate’, we add a fourth dimension.

The feedback for this talk was very strong and I have a couple good ideas moving forward, if only I can get some time to get a bit of code working.

Plane Home
One last little note. On the flight home, I was able to get the Cloudalicio.us engine to parse and process data from Connotea in addition to the native del.icio.us tagging sets I’ve been using. This means there’s hope for multiple parsers to be designed/coded in the wild. Cloudy may have its day, yet!

Tag Decay Poster from ASIST is online

Terrell Russell — Wed, 07 Nov 2007 19:12:26 +0000

I met quite a few people at the recent ASIS&T Annual Meeting in Milwaukee and told them I’d be getting my poster online.

The poster is up – Tag Decay: A View Into Aging Folksonomies (PDF 1.7MB)

It was a great problem to have people standing, listening, and asking questions for four hours. I just wish my throat had been warned ahead of time – I didn’t talk much the next day.

Thanks everyone.

Cloudalicio.us is back online!

Terrell Russell — Thu, 16 Nov 2006 05:13:05 +0000

So this is fun. Today, I realized I could probably get Cloudalicio.us back online and functional (it’s been offline since March due to a lack of data to parse and graph). It turns out, it was only a few lines of code to tweak and upload.

And it works – Meta Cloudy

–

The full post is here:

In preparing slides for my talk earlier today (770kb pdf), I was grabbing some screenshots from del.icio.us and rummaging through some of my old Cloudalicio.us graphs from before we went offline in March.

In clicking around at del.icio.us, I realized that the /url history pages are making available (again) the full post history of the users who have bookmarked that URL.

Before the shutdown in March, the /url history pages were displaying full historical posting data, providing the three entities that make up folksonomy (person, item, tags) as well as the date on which the tagging occurred. Cloudalicio.us would grab that data, parse it, and then display a graph allowing the visitor to see how the del.icio.us-using-public’s opinion of that page had changed over time.

Today, I realized that the full folksonomy data was again available (person, item, tags) but the date data is now compressed into postings by month. And that’s okay – it’s actually less points to graph, so the graphs should generate a little more quickly now. Additionally, I’ve set the cache to 15 days instead of 1 day as daily updates wouldn’t look that different on these new graphs anyways.

So, Behold, we’re back! Everybody say “Welcome back, Cloudy”!

Oh, and I’ve updated the greasemonkey script (v3) as well.