Wednesday, March 7, 2007

The Social Map Project

The Social Map Project is something I'm pursing along side Wantsy. Since this is really a blog about the life of an entrepreneur and techie, I'll also post information here.

After taking a look at Geni, it suddenly hit me that mapping out a user's known connections is very important. While services like Facebook and MySpace focus on the facilitation of "fake friends", I think it's very important to define a user's real social network.

How can this be done? With the use of existing data -- very primitive social networks like e-mail and IM "buddylists". The vast majority of people on the internet have an e-mail address and use it to communicate with their friends and family. For those that do not, they most likely use AIM. The data flow doesn't stop there, either. Most people are willing to trust someone that lives geographically close (i.e. a few doors down) than some random person 200 miles away.

I would really like to map out everyone's real social network using this primitive data. The barrier would have to be really low -- no user registration, very little information, and input as simple as possible. For example, I would simply input my e-mail address and name, and then input my mom's e-mail address and her name. That's it -- I'm mapped with my first connection - my mom. An e-mail then goes to her asking if she knows me... If she does, then it's a mutual connection. If she doesn't (or she doesn't respond), she's still listed as my connection, but it's a one-way.

In the future, it might be possible to mine user connections. For example, if I input my address, my geographical neighbors would be mapped in the database as "neighbors". Going a step further, if people begin to use meta-data as theorized for the "semantic web", you would be able to define connections by your interactions on other services or websites. For example, I interact with a ton of people in blogs... They would be more familiar to me than someone that I've never come across (although, the trust level would certainly be lower than immediate friends and family).

The concept isn't really a service at all, but rather a database of connections. I hate to use the term "social networking" because there really isn't any "working" going on at user level. Don't be fooled, however. There are many extremely valuable uses for such a database:
  1. User Trust Algorithm -- This could be used when no other metric is available. A user would simple input their e-mail address into the widget and see how they are connected with the user. Obviously a user in the 6th degree has more trust than someone in the 10th degree, but it doesn't stop there. If a user has lots of mutual connections several degrees out, then they are a trusted user. If they are simply an isolated node in shadowy corner, then they aren't that trusted.
  2. Intelligent E-mail -- This is pretty similar to the above. If people send you e-mail, it would get a rating dependent on how close a connection is. Spammers are very unlikely to be within your first several degrees (or have any meaningful mutual connections beyond on the second degree, for that matter).
  3. Data Sharing -- With a web built around trust, buying and selling would be safer, distributing information would be easier, etc..

6 comments:

Kevin said...

Cool idea, man. Geography is such an important part of social networking, I think it's a key feature for something like this (similar to the stuff Steve is talking about on Techquila Shots today). Especially if you're using it as a dating service - some long distance stuff works out online, but I'd much rather invest my time in someone I can meet for coffee without dropping $400 on a place ticket to Topeka. Gotta make it fun, though! :)

Robert said...

While user geography could definitely be plotted on a "map", my reference was hinting strongly towards a sociological map (or "tree")...

I want to create a web of trust. Trust can be derived from a connections database. For example, a spammer wouldn't climb very far in the system, as 1) nobody wants to connect to him, and 2) even if he managed to forge mutual connections, he would be isolated and his network wouldn't extend far out.

People like trust, and people like seeing who is connected to them. If people can understand the value of a user-to-user relational database (and I'm sure they will), the project should thrive.

The system will be open, allowing other services to utilize the database to extract relationship information while keeping everything private. An example would be Gmail implementing an algorithm to prevent spam... They would be able to define your connections many degrees out and assign a priority to your mail; all without seeing you or your friend's private information.

TJ said...

I'm working on something similiar, but in a more direct manner. I would love to see this come to life to incorporate it.

When do you forsee a an alpha/beta launch?

Robert said...

Hey TJ!

Right now we're looking at a closed alpha/beta around late May.

We're currently working on the database and working on methods for mining user-to-user relationship data.

Kiran Mova said...

Nice idea Robert.

I have been thinking in the same lines as you mention, social networking isn't really 'working' yet. The real value would be when people can move their real world relationships online and use that as a medium. I am working on such an system along with few other friends.

One of the concerns was to keep this network protected from spammers and I like your idea of building the network using links, and ranking the user with lot of one way links as a spammer. Would love to see your idea implemented and use it.

George Black said...

Very interesting to read your post. We have created a social networking site which, as you say, isn't really social networking as it is just for the people you already know and do things with. Not your 1000 Facebook friends.

I would love to talk to you more as we have a strong overlap in outlook and your data miner might of great use to us.