Since I have access to use it to figure out the influential voices online shaping the debate about John McCain’s choice of Sarah Palin as Republican VP Candidate last Friday  - here’s the list of “influentials” though you’ll need Radian6 to pull up the actual blog posts that are being listed) - the file is called sp-2.

And here’s what it looks like in Radian6

However Radian6 lacks something I think they need to add - as it becomes apperant when you try to drill down to find out sentiment for the Palin choice - for me there’s no easy way to create a set of comparision phases that accurately define the current opinions online - nor are Search Engines any help, since they just pick sort and rank phrases, not the sentiment around them.

Also, people often don’t say what they mean - when you’ve got millions of people repeating slang and common terms, it’s often difficult to settle on any set of terms that accurately defines opinion.

Another thing I wish Radian6 had, was a “nearness” or “close to” operator, similar to what Search Engines have had for a while, particularly Google - but most do have it - it looks for how close a word or phase is to another and makes a determination of relevence of the two phrases by their “proximity” to each other - there’s nothing in Radian6 today, that does that.  They should add it - but I imagine, it will add to processing time and make calculations of engagement and relevance more complicated.

Still, Radian6 and other tools like it need “Proximity” operations - saying “inexerienced” directly next to Sarah Palin’s name is different than saying “inexperienced” a few paragraphs away from her name.  In this way, Radian6 needs to incorporate some of Google’s intelligence - but that’s just my opinion - otherwise, it’s too much work to wade through every post to figure out what the sentiment and point of it is.

Another example - look at this - Scary - is Sarah Palin going to end up being US President? Somewhere else Hillary Is Smiling

But when you look at the actual posts in the “experience” category I created, you see its’ all over the spectrum - in other words, I could not get down to a filter that was sufficently precise to just those posts that expressed Sarah Palin had enough experience against those that said she didn’t have enough experience.

What would be great is if Radian6 did that for me.  Right …. Take those posts from Blogs, Mainstream Media and break them into phrase combinations - and then - offer to run those through a comparative topic monitor (shown above) for you - that way, at least some of the work of wading through conflicting opinions will be done for us.

And, another source of possible ideas for how to pick the phases you’d run into the comparative topic monitor (let alone Sentiment Analysis - which really isn’t done fully, or enough, today) is tapping into the BlogPulse Key Phrase Analysis for the day - here’s part of the analysis for August 30th, 2008 pertaining to the Sarah Palin VP Choice, the latest one at the time of me writing this post:

If we could get even to the level of content analysis of isolating phases like those above (”disaffected hillary clinton supporters” AND “sarah palin” with “good choice” or “bad choice” or something that works) we’d magnify what tools like Radian6, already very powerful, can do.

Right now, I could probably take those phrases and stick them into the Comparative Topic Monitor - no big deal - so I did that - at least for the top 6 phases - there were quite a few more but this is enough to prove a point.

Of the terms I picked “Palin pick” occured the most often in the last two days - and finally, I get to something I can use today -and can chart over time:

Look at this:

Now we have something - still very crude - but useful - perhaps, even a KPI metric we can monitor over time (were I in the political monitoring business - which I was some time - it feels like “horse racing” - sorta - handicapping the winners - too bad there’s no prize - like a million dollar ticket on a bet - oh well - just wishful thinking.

But here, at least, in the chart above, after a 4 or 5 hours of work - I managed to figure out some way to at least “filter” what Radian6 supplies, out of the box, into a “Yes” / “No” metric.

Of the content that had “Palin pick” in the last two days, about 1/3 also had the phase (experienced or inexperienced) - I filtered those out and today - it looks like twice as many voices (conversations) think Sarah Palin is too inexperienced to be Vice President than those who think she should - again, a  useful KPI (were I John McCain or Barack Obama, or running those campaigns).   If that number gets a lot smaller - if say 20% think Sarah Palin is not experienced enough in 2 months time, vs what it is today - I’d day, that’s very bad news for the Democrats.

But even here - if we ffurther refine this view of who thinks Sarah Palin is too inexperienced to be Vice President - to the top 10 influencers - also offered by Radian 6

.. there’s only one place where the debate is really being shaped, and that’s the Penny Arcade (ha!)

Whereas, for the view that Sarah Palin is experienced enough - we get three top sources where that debate is being shaped:

A lot of it (opinion shaping on weather Sarah Palin has the experience) is on the AOL Message Boards, Go My Town, and Campaign Diaries - I took an except from Campaign Diaries post, below

“…If my first post’s general sense was that McCain had made a strong move, this one is going in the opposite direction. My indecision isn’t surprising: This is a huge gamble and we will have to see how Palin performs in the months ahead, what coverage she receives and whether she can move the female vote before having a better idea of how her presence on the Republican ticket will play out.”

So the metrics, if one wants to pull out of this last statement - for how this all will play out - in some measure, might now be how well “Palin performs” and if the “female vote” of “dissaffected Hillary Supporters” actually moves over to Palin, or not - or how much it does move over to Palin.

This - the movement of the “dissaffected Hillary Supports” would be the main metric I’d suggest monitoring - via polling and Radian6 - provided you can get a decent filtering mechinism into Radian6 that pulls out the noise in the content being crawled - again, not an easy thing to do.

Right now - I’m not too happy with the map I’m seeing below - the electoral collage as it appears today - I think it’s way too close to call - and that scares me -

Let’s hope Barack Obama will use Hillary Clinton to highlight how different they both really are - it’s not just about woman in the White House - its what kind person is there - what their values are - let’s hope - because this map, is scary - I don’t care what anyone else is saying - thre’s not much Blue in the map - way too much Red.

Not good for Barack Obama - he needs to get a lot more Blue in that map -that’s my take on August 31st, 2008.