What I learned about Social Search on Yi-Tan and How It Applies to NpTech Tag Discussion ....

That is one visualization of the nptech data from Chris Blow

Last week, we had a NpTech Conference call (see Marnie's notes here and my stabb at a summary here)  During the call Nancy White mentioned "social search" and later she forwarded this note about a Yi-Tan tech community call on the topic today.  So, I decided to participate in the half-hour call.  I learned a lot and just beginning to think about how it applies to the NpTech tag discussion.  So, this is a little bit raw.

My notes in no particular order:

The call started with a definition and some tools

Social Search is: people helping people find stuff.

• a social searching program called lejit
• Digg Swarm - visualization of what is going through Digg
• Cha Cha - a search engine that lets you chat with someone - partial human assisted search
• Google reader has google reader sharing - reblogging - Imagine a power RSS user culls posts that are really useful and clicks share - and creates an exhaust feed of items. 
• Some other urls here and here

The first perspective we heard was that folks are sharing via email because it works best.  They know and trust the source.  The existing social searches are simply too big for trust - the information flow to huge and overwhelming.

There was discussion about do you "scale" trust - (via social networking network).  There was also a point about it is not just what you know, but who you know.  Also, that you may be SME in one topic, but a newbie in another.

I shared what we are doing with the NpTech Tag and the manual summary or what Jerry called "Human Filter."  Jerry noted that the job description or job role of human filter is something that we're going to see more of that type of job.   I remember that this was a prediction in David Shenk's book called Data Smog: Surviving the Information Glut.

On the call, it was noted "We trust someone who we know to summarize but  to scale it won't work because you loose trust.  It is the issue of having multiple people filtering multiple streams with the resulting stream being so large and unyielding.  There is also the issue of wow do we get there when there are people who want to catch our attention for profit not for edification?

I also mentioned the desire to have some collaborative filtering - like digg, perhaps using pligg as Allan Benamer and Marshall had suggested.

When I mentioned this, Jerry suggested I take a look at Dabbledb -- especially the screencast on analyzing digg data with it.   

After the call, set up an account and tried grabbing the RSS feeds from del.icio.us nptech tag, the feedburner feed, and a few other sources.   Then I remembered the data that Chris Blow had extracted from NpTech and I wondered if I could grab that?  Then I found Chris's most recent experiments with the Many Eyes and the NpTech data.   I followed the links to Chris's account
here and managed to grab some of the data into spreadsheet and pull it into dabbledb. 

I love what Chris did with the data -- I love that IBM tool too.  I'd love to be able to extract a summary of that data using some of Davie Hyerle's concepts and software.

What this left me wondering was -- will we always confront the issue of  too much information - even with collaborative filtering?  When do we reach the point when a human brain is not needed - but agents are?

Comments

[Marnie Webb]

The information glut is the opposite side of reducing the powers of traditional information gatekeepers. Now we've got to figure out who/what the gatekeepers are.

I think that we're going to ask for editors -- that's what you're doing w/ the nptech tag after all -- and we're going to figure out how their sources are more exposed for the people who want to do the digging.