Let's call this: the SMA: Social Media Algebra. - directeur

sounds like Semantic Web to me. - MikeAmundsen

Mike, indeed. But the "formalism" is needed. Theories could be built. - directeur

People have APML profiles, but does content? - Jason Wehmhoener

Jason, good question! Yes, lots of content already have tagging information. Blog posts, bookmarking sites, music (last.fm), videos... - directeur

What I mean is that we really can build a great formalism for this. In such a way that we'll be able to work like mathematicians. Be able to prove stuff, discuss logically... (memories of my language theories, algebra, and other courses :) ) <parenthitcal statement with emoticon http://xkcd.com/541/ :) > - directeur

I would like to see more Social Media visualization of services and connections . . . SMAV - Chris Loft

Chris that'd be a nice application of the SMA. Don't we already have visualisation for sets, vectors, and other algebra things? :) - directeur

Would Social Media Mathematics be a more apt term for this proposed field, with Social Graph and SMA being categories? - coldbrew

coldbrew: aha! not bad, not bad at all! :) - directeur

what's 'algebra'? like something Ariel from the Little Mermaid would wear when she starts developing? - Morgan Haley

So, before you can have a proof, you need some postulates right? - Jason Wehmhoener

Jason, sure! We need some axioms... - directeur

I just discovered APML. Awesome. - Onur Gündüz

Onur: here's yours: http://www.socialwhois.com/ogunduz... :) - directeur

delicious :) reading specs right now. http://www.apml.org/develop... - Onur Gündüz

I'm stumped on axioms. What's a straight line in social media? What are some basic things we can say without ambiguity? - Jason Wehmhoener

Brian Solis posted a link to this post which attempts to use mathematics to describe, "a set of new measures based on ego networks": http://www.analytictech.com/connect... - coldbrew

It is taken from Structural Holes, Ron Burt (1992; 1995), and might be an interesting place to start. - coldbrew

coldbrew, funny just got an email from Brian this morning :) - directeur

i like your thinking @directeur. - .LAG liked that

Thanks .LAG! :) - directeur

I just looked over the Structural Holes paper. I've seen some other attempts at applying graph theory to networks, and it's true that you can discover some things about redundancy, density, etc, but I don't think any of that stuff really means anything useful until you throw APML into the mix. - Jason Wehmhoener

The large social networks have a shit-ton of APML-like data. It is clear that they are also looking at this topic intensely, in addition to sentiment analysis. - coldbrew

Simple arithmetic applied on interests, likes and dislikes should be sufficient to calculate matches of unlimited types among billions of people, organizations, groups and social networks worldwide. And ultra-simple semantic markup should be sufficient to specify those interests, likes and dislikes. No algebra required. - Sean McBride

Best would be to set all this attention-data free, so anyone can hack away on it... The companies that sit on this data now don't have the resources to extract all value. I think they're only scratching the surface on what we could derive. That's why more and more companies are opening their databases with an API, it enables mashups, innovation from outside the corporate borders. - Meryn Stol

Sean, nothing is required, but everything's welcome :) Imagine these concepts with me: social distance, social similarity, vectors algebra has nice stuff... I bet we can study many social media behaviors with a tool like the SMA. Don't you think? - directeur

directeur -- we could start using simple markup right now to express our machine-readable and scrapable interests, likes, dislikes and preferences in Friendfeed messages and comments. We simply need to agree on a markup standard to get the ball rolling. The more minimal the better. Friendfeeders have already expressed tens of thousands of sentiment data points on Friendfeed, but the content is inert, locked up in unstructured and unscrapable natural language text. Once the data points are universally formalized, then we can go to town on them with simple arithmetic or higher mathematics. - Sean McBride

This is good, I like the ideas. I've seen the term 'social algebra' used by teachers and social workers. Mathematically, this may be an abstract algebra, but it certainly would qualify as group/ring/field theory. (I'm not sure I understand enough about it, ie the functions that we define would have to have commutitivity, associativity, identity functions (0 & 1), inverse functions) - Greg GuitarBuster

"the content is inert, locked up in unstructured and unscrapable natural language text" I don't agree with that. There's even a Twitter app which gauges sentiments on subjects based on natural language. It won't be flawless, and it won't capture anything, but natural language is far from impossible to work with. Simple regular expressions can be a begin, certainly if you don't mind throwing away most of the data (because you can only successfully parse a small bit). - Meryn Stol

An algebra for social media must be based on an ontology of social relations. What are all the ways in which social objects can relate to and interact with one other? What is the best way to represent those relations and interactions in a formal markup language that is easy for everyone to use? - Sean McBride

THere are at least two sets of data not immediately related. 1) the social graph based on xfn and 2) the interests based on APML. They are not immediately mixable. - François Granger

To build an algebra, I would first define some simple purpose. Then starting with this the base of the algebra could be built. For example, on the APML side, we would like to know how close are peoples based on the similarity of their interests. we would also like to build sets of people like minded. On the social graph side, we could try to define closeness based on the number of services where two people are linked compared to the total number of services where they are both present - François Granger

for exemle using a service like http://www.usernamecheck.com/ - François Granger